Molecular Dissection of a Borrelia burgdorferi In Vivo Essential PurineTransport System

Sunny Jain, Adrienne C. Showman, Mollie W. Jewett

Burnett School of Biomedical Sciences, University of Central Florida College of Medicine, Orlando, Florida, USA

The Lyme disease spirochete Borrelia burgdorferi is dependent on purine salvage from the host environment for survival. Thegenes bbb22 and bbb23 encode purine permeases that are essential for B. burgdorferi mouse infectivity. We now demonstrate theunique contributions of each of these genes to purine transport and murine infection. The affinities of spirochetes carryingbbb22 alone for hypoxanthine and adenine were similar to those of spirochetes carrying both genes. Spirochetes carrying bbb22alone were able to achieve wild-type levels of adenine saturation but not hypoxanthine saturation, suggesting that maximal hy-poxanthine uptake requires the presence of bbb23. Moreover, the purine transport activity conferred by bbb22 was dependent onan additional distal transcriptional start site located within the bbb23 open reading frame. The initial rates of uptake of hypoxan-thine and adenine by spirochetes carrying bbb23 alone were below the level of detection. However, these spirochetes demon-strated a measurable increase in hypoxanthine uptake over a 30-min time course. Our findings indicate that bbb22-dependentadenine transport is essential for B. burgdorferi survival in mice. The bbb23 gene was dispensable for B. burgdorferi mouse infec-tivity, yet its presence was required along with that of bbb22 for B. burgdorferi to achieve maximal spirochete loads in infectedmouse tissues. These data demonstrate that both genes, bbb22 and bbb23, are critical for B. burgdorferi to achieve wild-type in-fection of mice and that the differences in the capabilities of the two transporters may reflect distinct purine salvage needs thatthe spirochete encounters throughout its natural infectious cycle.

Purine nucleobases are required for the synthesis of DNA andRNA. Consequently, purine biosynthesis and/or transport is a

critical process across all kingdoms of life. Moreover, nucleobasetransporters represent possible therapeutic targets for cancer andinfectious diseases (1).

Borrelia burgdorferi, the causative agent of Lyme disease, has alimited genome and lacks the enzymes required for de novo purinebiosynthesis (2–6). Therefore, purine salvage from host environ-ments is important for B. burgdorferi survival and pathogenesis(7–9). Recently, our laboratory established that the genes bbb22and bbb23, located on B. burgdorferi essential circular plasmid 26(cp26), together encode a purine transport system that is requiredfor hypoxanthine transport and contributes to adenine and gua-nine transport (7). Spirochetes lacking both bbb22 and bbb23 arenoninfectious in mice (7). Conversely, these genes are dispensablefor B. burgdorferi growth in nutrient-rich medium in vitro (7).

The bbb22 and bbb23 genes encode proteins within clusterCOG2252 of the nucleobase cation symporter-2 superfamily(NCS2) (7), which includes permeases found in bacteria, fungi,and plants that are specific for adenine, hypoxanthine, and/orguanine (10–16). The Aspergillus nidulans AzgA transporter, thefounding member of this family of transporters, has specificity foradenine, hypoxanthine, and guanine (10, 11, 13). In other species,however, specific purine transport functions have been attributedto distinct proteins. For example, Escherichia coli harbors twohigh-affinity adenine permeases, PurP and YicO, and two high-affinity hypoxanthine-guanine permeases, YjcD and YgfQ (12,15). Similarly, Bacillus subtilis has distinct uptake systems for hy-poxanthine-guanine and adenine (16, 17). The Arabidopsis thali-ana proteins Azg1 (AtAzg1) and AtAzg2 have been shown to beplant adenine-guanine transporters (14) although hypoxanthineuptake has yet to be examined for these proteins. The B. burgdor-feri BBB22 and BBB23 open reading frames (ORFs) share 79.8%and 78.3% identity at the nucleic acid and amino acid levels, re-

spectively, and are adjacent genes on cp26. This suggests that thetwo genes may be paralogs (7). Although the transport activitiesconferred by bbb22 and bbb23 together closely resemble the ade-nine-hypoxanthine-guanine permease function of AzgA (7, 10,13), the individual contributions of bbb22 and bbb23 to the uptakeof specific purines remain unknown. Although the overall aminoacid sequences of the BBB22 and BBB23 proteins are highly con-served, sufficient sequence divergence may occur within two ofthe six putative periplasmic loops, which may allow for differentialsubstrate specificity between the two proteins. In addition, it ispossible that the two genes may confer the same transport func-tions but are differentially expressed (7).

Here, we analyzed the individual roles of bbb22 and bbb23 inpurine transport and murine infection. Our findings demon-strated that BBB22 and BBB23 were each capable of hypoxan-thine, adenine, and guanine uptake, albeit with differing affinities.Moreover, bbb22, but not bbb23, alone restored murine infectivityto �bbb22-bbb23 (�bbb22-23) spirochetes. Nonetheless, the spi-rochetes carrying bbb22 alone were unable to achieve spirocheteloads in the infected mouse tissues equivalent to the bbb22-23�

Received 31 October 2014 Returned for modification 19 November 2014Accepted 10 March 2015

Accepted manuscript posted online 16 March 2015

Citation Jain S, Showman AC, Jewett MW. 2015. Molecular dissection of a Borreliaburgdorferi in vivo essential purine transport system. Infect Immun 83:2224 –2233.doi:10.1128/IAI.02859-14.

Editor: R. P. Morrison

Address correspondence to Mollie W. Jewett, Mollie.Jewett@ucf.edu.

S.J. and A.C.S. contributed equally to this article.

Copyright © 2015, American Society for Microbiology. All Rights Reserved.

doi:10.1128/IAI.02859-14

2224 iai.asm.org June 2015 Volume 83 Number 6Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

spirochete loads. Together, these data suggest that both bbb22 andbbb23 are critical for B. burgdorferi to achieve wild-type levels ofpurine transport and mouse infection.

MATERIALS AND METHODSBacterial clones and growth conditions. Escherichia coli strain DH5� wasgrown in LB broth or on LB agar plates at 37°C. Gentamicin and specti-nomycin were used at 10 �g/ml and 300 �g/ml, respectively. All B. burg-dorferi clones were generated in the low-passage-number B31 A3-68-�BBE02-�bbb22-23 clone, which lacks genes bbb22-23 and bbe02 andplasmids cp9 and linear plasmid 56 (lp56) (7). B. burgdorferi clones weregrown in Barbour-Stoenner-Kelly II (BSKII) medium at 35°C. Gentami-cin, streptomycin, and kanamycin were used at 40 �g/ml, 50 �g/ml, and200 �g/ml, respectively.

5= RACE. Total RNA was isolated with TRIzol reagent (Life Technol-ogies) from a 50-ml culture of B. burgdorferi clone B31 A3 grown in BSKIImedium to a density of 4 � 107 spirochetes/ml (log phase) or from a 5-mlculture at a density 2 � 108 spirochetes/ml (stationary phase). 5= Rapidamplification of cDNA ends (5=RACE) using a 5=RACE System for RapidAmplification of cDNA Ends, version 2.0 (Life Technologies), was per-formed according to the manufacturer’s instructions. Four micrograms ofRNA was used per reaction volume along with gene-specific primers 1297,1298, and 1299 for bbb22, primers 1316, 1317, and 1318 for the transcript

within bbb23, and primers 1300, 1301, and 1302 for bbb23 (Table 1).Nested PCR products were purified using a PCR purification kit (Qiagen),and the 5= end of each transcript was determined by sequence analysis(Genewiz).

Generation of plasmids containing bbb22 and bbb23. A sequenceregion of 107 bp upstream of the BBB23 ORF containing the bbb23 pro-moter (23p) (7) and flaB promoter sequence were amplified from B. burg-dorferi B31 clone A3 genomic DNA using Phusion enzyme (Thermo Sci-entific) and the primer pair 1526 and 1527 and the pair 1577 and 1578(Table 1), respectively. The B. burgdorferi shuttle vector pBSV2G (18) andpromoter fragments were digested with restriction enzymes KpnI andBamHI, ligated, and cloned in E. coli. Plasmids were confirmed by restric-tion digestion and DNA sequence analysis. The annotated BBB23 ORFsequence was amplified from B31 A3 genomic DNA using Phusion en-zyme (Thermo Scientific) and the primer pair 1006 and 1525. A subse-quent PCR using the primer pair 1006 and 1011 (Table 1) and the PCRproduct derived from primer pair 1006 and 1525 (Table 1) as a templatewas used to replace the 3=HindIII restriction site with SalI. The bbb23 genewas ligated into pBSV2G 23p and pBSV2G flaBp plasmids using restrictionenzymes BamHI and SalI and cloned in E. coli. Plasmids pBSV2G 23p-bbb23� and pBSV2G flaBp-bbb23� were confirmed by restriction diges-tion and DNA sequence analysis. DNA fragments containing the bbb22ORF and 1,466 bp of upstream sequence (bbb22 distal promoter, 22distp)

TABLE 1 Primers and probes used in this study

Primer no. Primer name Sequence (5=–3=)a

1006 BBB23 ORF 1392-BamHI-Fwd CCCGGGGGATCCTTGAAAATATTTTTAAACAATAAAAAGGAAAGTTT1007 BBB23 ORF 1356-Fwd-BamHI CCCGGGGGATCCATGGGTAAGTATGTAAAAGGTTTATTTTTT1011 BBB23 ORF 1392-Rev-SalI CCCGGGGTCGACTTACAGATCTTCTTCAGAAATAAGTTTTTGTTCATAGCC1123 recA-F AATAAGGATGAGGATTGGTG1124 recA-R GAACCTCAAGTCTAAGAGATG1137 flaB-TaqMan-Fwd TCTTTTCTCTGGTGAGGGAGCT1037 pUC18F CCCAGTCACGACGTTGTAAAAC1138 flaB-TaqMan-Rev TCCTTCCTGTTGAACACCCTCT1140 nid-TaqMan-Fwd CACCCAGCTTCGGCTCAGTA1141 nid-TaqMan-Rev TCCCCAGGCCATCGGT1166 BBB23 TaqMan Fwd GCACTAGTTACCGCAACCTGTCT1167 BBB23 TaqMan Rev CTCATTCCAGAAGCCAATGCT1168 BBB22 TaqMan Fwd GAATCAATCCAAAGAAACATTGTTAT1169 BBB22 TaqMan Rev AGTTGCAGTAACTAATGCGCCAAT1297 BBB22 GSP-1A TACTGCGCTTTCTGGAC1298 BBB22 GSP-1B AATTTGCAATACTTTCCCTAGC1299 BBB22 GSP-1D GCTGATGTTAAGCAAGTTGCAG1300 BBB23 GSP-1A CCTGGGCACCTTCTAAAT1301 BBB23 GSP-1B AGTTTATAATTTGCTCCCTTACTC1302 BBB23 GSP-1D TGCTGTTAGACAGGTTGCG1316 BBB22 IVET RACE GSP1 GGCAAAAAAAACTGCAAATAGAAATAATG1317 BBB22 IVET RACE GSP2 GCCGTCTACCAGTAATATTTTTTTTGC1318 BBB22 IVET RACE GSP3 ATTGAACAAGAGAATAAAAACTATTGATATAAAAGTCC1525 BBB23-cmyc-HindIII Rev CCCGGGAAGCTTTTACAGATCTTCTTCAGAAATAAGTTTTTGTTCATA

GCCA TAAATAAATTTAATAATAAAAATTAAGC1526 BBB23 promoter KpnI Fwd CCCGGGGGTACCTATTTTCAAACTTTACCTGACAGCG1527 BBB23 promoter BamHI Rev CCCGGGGGATCCTAATAATTATTAGGCTTTTAATTCTTTTTAAA1528 BBB22 BamHI-Fwd CGGGATCCTCTCCTGTACTGCTAATATTATGC1529 BBB22 HindIII- Rev CCCAAGCTTGACTCTTTTTTATGATTTATAATTTAGG1577 flaBp-fwd-KpnI CCCGGGGGTACCTGTCTGTCGCCTCTTGTGGCT1578 flaBp-Rev-BamHI GGGGGATCCGATTGATAATCATATATCATTCCT1668 pUC18R AGCGGATAACAATTTCACACAG1139 flaB TaqMan probe 6-FAM-AAACTGCTCAGGCTGCACCGGTTC-TAMRA1142 nid TaqMan probe 6-FAM-CGCCTTTCCTGGCTGACTTGGACA-TAMRA1575 BBB22 TaqMan probe 6-FAM-AGCATGGCATACATAATAGCTGTTAATCCGGC-TAMRA1576 BBB23 TaqMan probe 6-FAM-CTAATGGGACTTTATACCAATACGCCTT-TAMRAa FAM, 6-carboxyfluorescein; TAMRA, 6-carboxytetramethylrhodamine. Restriction sites are in boldface.

BBB22 and BBB23 Transport Function

June 2015 Volume 83 Number 6 iai.asm.org 2225Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

or the bbb22 ORF and 160 bp of upstream sequence (bbb22 proximalpromoter, 22proxp) were amplified from B31 A3 genomic DNA usingPhusion enzyme (Thermo Scientific) and the primer pair 1007 and 1529and the pair 1528 and 1529, respectively (Table 1). The DNA fragmentswere ligated into the pBSV2G vector using BamHI and HindIII restrictionsites and the plasmids were cloned in E. coli. Plasmids pBSV2G 22distp-bbb22� and pBSV2G 22proxp-bbb22� were confirmed by restriction di-gestion and DNA sequence analysis (Genewiz).

Reintroduction of bbb22 and bbb23 into bbb22-23 mutant spiro-chetes. B. burgdorferi clone B31 A3-68-�BBE02-�bbb22-23 was trans-formed with 20 �g of plasmid pBSV2G 23p-bbb23�, pBSV2G flaBp-bbb23�, pBSV2G 22proxp-bbb22�, or pBSV2G 22distp-bbb22�, and thetransformants were selected in BSK1.5X semisolid medium containing1.7% agarose and 40 �g/ml gentamicin at 37°C in 2.5% CO2 for 6 to 8days. B. burgdorferi transformants were confirmed by colony PCR usingprimers 1037 and 1668 (Table 1). The endogenous plasmid content ofeach new B. burgdorferi clone was confirmed by PCR analysis (8) to be thesame as that of the B31 A3-68-�bbe02-�bbb22-23 parent clone (7) (Table 2).

RNA extraction from in vitro-grown cultures. B. burgdorferi clonesbbb22-23�, �bbb22-23, 23p-bbb23�, flaBp-bbb23�, 22proxp-bbb22�, and22distp-bbb22� (Table 2) were grown to a density of 2 � 108 spiro-chetes/ml in BSKII medium containing the appropriate antibiotics. A to-tal of 2 � 108 spirochetes were harvested at 5,800 � g for 10 min. RNA wasisolated using TRIzol reagent (Life Technologies) or a Direct-zol RNAminiprep kit (Epigenetics) and resuspended in 50 �l of diethyl pyrocar-bonate (DEPC)-treated distilled H2O (dH2O). RNA samples were DNasetreated using Turbo DNA-free (Life Technologies) and were confirmed tobe free of B. burgdorferi genomic DNA contamination by PCR using prim-ers 1123 and 1124. One microliter of SUPERaseIn RNase inhibitor (LifeTechnologies) was added to prevent RNA degradation, and samples werestored at �80°C.

Gene expression analysis. A total of 400 ng of B. burgdorferi RNA,extracted from late stationary-phase spirochetes as described above, wasused to synthesize cDNA with an iScript Select cDNA synthesis kit (Bio-Rad) using random hexamer primers. Reaction mixtures lacking reversetranscriptase were run in tandem as negative controls. TaqMan real-timequantitative PCR (qPCR) was performed using 400 ng of cDNA, IQ SuperMix (Bio-Rad), and the following primer/probe sets: 1168, 1169, and 1575(bbb22); 1166, 1167, and 1576 (bbb23); and 1137, 1138, and 1139 (flaB)(Table 1). mRNA transcript copy numbers for bbb22 and bbb23 werenormalized to flaB mRNA copy numbers. All data sets were comparedusing one-way analysis of variance (ANOVA), followed by Tukey’s post-test (GraphPad Prism, version 5.0).

Radioactive transport assays. B. burgdorferi clones were grown to astate of late stationary-phase starvation at a density of 2 � 108 spirochetes/ml. Radioactive transport assays were performed as previously described(7). Briefly, to initiate the metabolism of the starved spirochetes, the bac-teria were preincubated at 37°C for 30 min in 1-ml reaction volumescontaining HN buffer (50 mM HEPES and 50 mM NaCl, pH 7.4), 6 mMglucose, and 1.4% rabbit serum (Pelfreez). This amount of rabbit serumwas estimated to contain �90 nM hypoxanthine and �5 nM adenine (19,20), both of which represented less than 10% of the lowest concentrationsof [3H]hypoxanthine and [3H]adenine used in the kinetic assays de-

scribed below. Thirty-minute uptake experiments were performed using2.5 �M [3H]hypoxanthine monohydrochloride (specific activity, 20 Ci/mmol) (PerkinElmer) or 0.396 �M [2,8-3H]adenine (specific activity, 15Ci/mmol) (PerkinElmer), and samples were analyzed at 10 s and 2, 5, 8,15, and 30 min following the addition of radioactivity. The amounts ofhypoxanthine and adenine uptake were converted to femtomoles/5 � 107

spirochetes as previously described (7). Initial rate experiments were per-formed using 1.0 to 2.5 �M [3H]hypoxanthine monohydrochloride (spe-cific activity, 20 Ci/mmol) (PerkinElmer) or 0.066 to 0.396 �M [2,8-3H]adenine (specific activity, 15 Ci/mmol) (PerkinElmer), and sampleswere analyzed at 10, 20, 30, 45, and 60 s following the addition of radio-activity. The rates of purine uptake (fmol/5 � 107 spirochetes/s) wereestimated for each purine concentration by linear regression, and the Km

values were determined using GraphPad Prism, version 5.0. Data setswere compared using an unpaired t test or two-way ANOVA using Graph-Pad Prism, version 5.0.

For the [3H]hypoxanthine uptake assays in the presence of cold com-petitors, B. burgdorferi clones were grown and prepared as describedabove. Stock solutions (1 mM) of each cold competitor, hypoxanthine,guanine, adenine, and cytosine, were prepared in 8 mM NaOH. All thereactions were carried out in a 1-ml volume containing 1 � 109 spiro-chetes, 250 nM [3H]hypoxanthine, and either 5 �l of cold competitor (5�M final concentration, 20-fold excess) or 5 �l of 8 mM NaOH alone (40�M final concentration). This final concentration of NaOH in the reac-tion mixture was not found to affect the ability of the spirochetes to trans-port [3H]hypoxanthine (data not shown). Aliquots (100 �l) containing108 spirochetes were removed 15 min following the addition of radioac-tivity and were analyzed as previously described (7). The percent [3H]hy-poxanthine uptake for each reaction condition was calculated relative tothe amount of [3H]hypoxanthine transport in the absence of competitor.bbb22-23� spirochetes (Table 2) heated to 95°C for 15 min (heat killed)served as the negative control for all transport experiments.

Mouse infection experiments. The University of Central Florida(UCF) is accredited by International Association for Assessment and Ac-creditation of Laboratory Animal Care. All the mouse experiments weredone according to the guidelines of National Institutes of Health (NIH)and approved by the Institutional Animal Care and Use Committee ofUCF. Groups of 6- to 8-week-old C3H/HeN female mice (Harlan) wereinoculated with B. burgdorferi clone bbb22-23�, �bbb22-23, 23p-bbb23�,flaBp-bbb23�, 22proxp-bbb22�, or 22distp-bbb22� (Table 2). Mice wereinoculated intraperitoneally (80%) and subcutaneously (20%) with a to-tal dose of 1 � 104 or 1 � 107 spirochetes. Inoculum cultures were ana-lyzed for the endogenous plasmid content by PCR and plated on solid BSKmedium, and individual colonies from each population were analyzed forthe presence of virulence plasmids lp25, lp28-1, and lp36 (9). All inocu-lum cultures carried the expected endogenous plasmid content (Table 2),and 70 to 100% of individual colonies from each clone were confirmed tocontain all three virulence plasmids. Blood for serological analysis wascollected prior to inoculation and at 3 weeks postinoculation. Mouse in-fection was determined by serology and spirochete reisolation from tis-sues as previously described (7).

Quantitation of spirochete loads in mouse tissues. Ear, heart, andjoint tissues were collected for DNA extraction from mice 3 weeks post-inoculation. DNA extraction was performed as previously described (8).Real-time quantitative PCR was performed using 100 ng of DNA frominfected mouse tissues. The number of B. burgdorferi genomes was deter-mined using the standard curve method to quantify the number of B.burgdorferi flaB copies using flaB primers (1137 and 1138) and probe(1139) (96% efficiency). The standard curve for B. burgdorferi DNA wasgenerated using 0.001 ng, 0.01 ng, 0.1 ng, 1.0 ng, and 10.0 ng of B. burg-dorferi genomic DNA. Our assay was found to have a limit of detection of5 to 10 B. burgdorferi genomes per reaction. The number of mouse ge-nomes was quantified using mouse nid primers (1140 and 1141) andprobe (1142) (100% efficiency) (Table 1). The standard curve for mouseDNA was generated using 5.5 � 103, 5.5 � 104, 5.5 � 105, and 5.5 � 106

TABLE 2 B. burgdorferi clones used in this study

B. burgdorferi clonea Shuttle vector Reference or source

�bbb22-23 pBSV2G 7bbb22-23� pBSV2G bbb22-23� 723p-bbb23� pBSV2G 23p-bbb23� This workflaBp-bbb23� pBSV2G flaBp-bbb23� This work22proxp-bbb22� pBSV2G 22proxp-bbb22� This work22distp-bbb22� pBSV2G 22distp-bbb22� This worka All of the clones had the following genotype: B31 A3-68 bbe02::flgBp-kan bbb22-23::flaBp-aadA cp9� lp56�.

Jain et al.

2226 iai.asm.org June 2015 Volume 83 Number 6Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

nid copies (8). All qPCRs were performed using IQ supermix (Bio-Rad),as previously described (8). The data were reported as the number of B.burgdorferi genomes/104 mouse genomes, and the data sets were com-pared using one-way ANOVA with Tukey’s posttest using GraphPadPrism, version 5.0.

RESULTSThe bbb22 and bbb23 genes are separate transcripts. We previ-ously demonstrated that genes bbb22-23, present on circular plas-mid 26 (cp26), encode purine permeases and together are essentialfor B. burgdorferi mouse infection (7). However, the individualcontributions of bbb22 and bbb23 to purine transport and mouseinfection remain unknown. The annotated bbb22 and bbb23 openreading frames are separated by a 109-bp intergenic region (4),suggesting that these two genes may be transcribed separately andtherefore may play distinct roles in the ability of B. burgdorferi tosalvage purines. Rapid amplification of cDNA ends (RACE) wasused to identify the 5= ends of the bbb22 and bbb23 transcripts. Thetranscription start site for bbb23 was found to be 260 bp upstreamof the annotated bbb23 open reading frame (ORF), which corre-sponds to nucleotide 21082 on cp26 and is 225 bp within the 3=end of bbb24 (Fig. 1). The transcription start site for bbb22(bbb22prox) was found to be 16 bp upstream of the annotatedbbb22 ORF at nucleotide 19338 on cp26 (Fig. 1). In addition, an invivo expression technology (IVET)-based genetic screen per-formed by our laboratory to identify B. burgdorferi sequences thatare expressed during murine infection (21) identified a candidatepromoter sequence (B. burgdorferi IVE fragment 103 [Bbive103],nucleotides 20322 to 20238 on cp26) within and in the same ori-entation to the bbb23 ORF. Consistent with this finding, 5= RACEanalysis using RNA isolated from in vitro-grown spirochetes vali-dated the presence of a transcription start site at nucleotide 20219,17 bp downstream of the Bbive103 sequence (bbb22dist) (Fig. 1),indicating that this promoter is active both in vitro and in vivo. Thedouble sequence observed in the sequence analysis of thebbb22dist 5= RACE PCR product (Fig. 1B) reflects detection ofboth the transcript derived from the putative distal promoter forbbb22 and the bbb23 transcript itself. Together, these data indicatethat bbb23 and bbb22 have distinct transcription start sites andthat an additional transcription start site is present 897 bp up-stream of bbb22 within bbb23.

A long 5=UTR within bbb23 is required for wild-type expres-sion of bbb22. Toward the goal of understanding the individualroles of bbb22 and bbb23, a panel of plasmids was engineered usingthe B. burgdorferi shuttle vector pBSV2G (18) to contain the bbb23ORF along with its endogenous promoter sequence (23p-bbb23�),the bbb23 ORF along with the constitutive flaB promoter (flaBp-bbb23�), the bbb22 ORF along with the proximal promoter se-quence (22proxp-bbb22�), or the bbb22 ORF along with the prox-imal and distal promoter sequences (22distp-bbb22�) (Table 2). Inorder to generate spirochetes that harbored only bbb22 or bbb23,the shuttle vector plasmids carrying bbb23 or bbb22 were eachtransformed into a low-passage-number B. burgdorferi clone thatlacks both bbb22 and bbb23 (7). All B. burgdorferi transformantswere verified to contain all of the endogenous plasmid repliconspresent in the �bbb22-bbb23 mutant parent (Table 2). The in vitrogene expression of bbb22 and/or bbb23 was analyzed in the B.burgdorferi clones bbb22-23�, 22proxp-bbb22�, 22distp-bbb22�,23p-bbb23�, and flaBp-bbb23� (Table 2). Each was grown to latestationary phase, a starvation condition that we have previously

demonstrated to induce purine permease activity (7) (Fig. 2). Thelevels of bbb22 and bbb23 expression in the bbb22-bbb23� clonewere equivalent to the levels of expression of these genes in wild-type spirochetes (data not shown). The expression level of bbb22

FIG 1 Genes bbb22 and bbb23 each harbor their own promoters. Rapid amplifi-cation of cDNA ends (5= RACE) analysis of the transcription start sites for thebbb23, bbb22prox, and bbb22dist transcripts was performed using RNA extractedfrom B31 clone A3 B. burgdorferi. (A) Transcript-specific 5= anchored nested PCRproducts generated from the 5= RACE analysis. Data shown are representative oftwo replicate experiments. The DNA ladder is shown in base pairs. (B) DNAsequence chromatographs of the DNA sequence of the 5= RACE products shownin panel A. Both sequences detected for bbb22dist are shown. The 5= nucleotide ofeach transcript is indicated with a star. (C) Schematic diagram of the genetic or-ganization of genes bbb22 and bbb23. Transcription start sites identified in panel Bare shown as bent arrows and labeled with the corresponding transcripts names.The B. burgdorferi IVET in vivo active promoter sequence within the BBB23 ORFis shown as a light gray box. The locations and numbers of the transcript-specificprimers used to generate the cDNA for the 5=RACE analysis are indicated. A rulerfor the nucleotide coordinates on cp26 is shown.

BBB22 and BBB23 Transport Function

June 2015 Volume 83 Number 6 iai.asm.org 2227Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

under the control of the proximal promoter alone (22proxp-bbb22�) was significantly reduced compared to the level of bbb22expression in the bbb22-23� positive-control spirochetes (Fig.2A). Interestingly, the level of bbb22 expression in 22distp-bbb22�

spirochetes was similar to that of bbb22-23� spirochetes (Fig. 2A).Together, these data suggest a possible role for the long 5=untrans-lated region (UTR) in the 22distp-bbb22� construct in the regula-tion of wild-type levels of bbb22. There was no significantdifference detected between the levels of bbb23 transcription in

23p-bbb23� spirochetes and levels in bbb22-23� spirochetes (Fig.2B), suggesting that the upstream sequence included in this con-struct was sufficient to drive bbb23 expression to wild-type levels.It was observed that in late stationary-phase growth, the level ofbbb23 expression in bbb22-23� spirochetes (Fig. 2B) was approx-imately 100-fold lower than the level of bbb22 expression in thesame clone (Fig. 2A). A similar, significant difference between thelevels of late stationary-phase expression of bbb22 and bbb23 wasdetected in wild-type spirochetes (data not shown). We found thatby driving expression of bbb23 under the control of the constitu-tive promoter flaB, the level of bbb23 expression in late stationaryphase could be increased to a level similar to that of bbb22 underthe control of its endogenous promoter (Fig. 2).

Both bbb22 and bbb23 are necessary for B. burgdorferi toachieve wild-type levels of hypoxanthine transport. We previ-ously demonstrated the combined roles of bbb22 and bbb23 inhypoxanthine transport (7). To determine the ability of spirochetesharboring either bbb22 or bbb23 alone to transport hypoxanthine,uptake assays were performed using a panel of B. burgdorferi clonescarrying either bbb22 or bbb23 (Table 2). Hypoxanthine uptakewas measured using 2.5 �M [3H]hypoxanthine over a 30-mintime course. This concentration of hypoxanthine is similar to thephysiological concentration of hypoxanthine in mammalianplasma (22). The amounts of [3H]hypoxanthine detected in thespirochetes carrying only bbb22 or bbb23 were significantly re-duced compared to the amounts in the bbb22-23� clone (Fig. 3A).

FIG 2 In vitro gene expression of bbb22 and bbb23 in B. burgdorferi clones.RNA was extracted from bbb22-23�, 23p-bbb23 �, flaBp-bbb23�, 22proxp-bbb22�, and 22distp-bbb22� B. burgdorferi clones at late stationary phase (2 �108 spirochetes/ml) grown at 37°C. Gene expression for bbb22, bbb23, and flaBwas quantified by reverse transcriptase qPCR. (A) bbb22 expression was ana-lyzed for clones bbb22-23�, 22proxp-bbb22�, and 22distp-bbb22�. bbb22mRNA transcripts were normalized to the number of flaB mRNA copies. (B)bbb23 expression was analyzed for clones bbb22-23�, 23p-bbb23�, and flaBp-bbb23�. bbb23 mRNA transcripts were normalized to the number of flaBmRNA copies. Data represent the average values of three biological replicates.Error bars represent the standard deviations from the means. The y axis isdepicted in a log10 scale. Data sets were compared by one-way ANOVA usingGraphPad Prism, version 5.0. *, P 0.05; **, P 0.01; NS, not significant.

FIG 3 bbb22 and bbb23 alone provide spirochetes with distinct abilities to takeup purines. Radioactive purine uptake by B. burgdorferi bbb22-23�, 23p-bbb23�, flaBp-bbb23�, 22proxp-bbb22�, 22distp-bbb22�, �bbb22-23, andbbb22-23� heat-killed spirochetes was measured over a 30-min time coursefollowing the addition of 2.5 �M [3H]hypoxanthine (A) or 0.396 �M [2,8-3H]adenine (B).The specific activity of the labeled hypoxanthine used was 20Ci/mmol. The specific activity of the labeled adenine used was 15 Ci/mmol.Error bars represent the standard deviation from the mean for at least twobiological replicates.

Jain et al.

2228 iai.asm.org June 2015 Volume 83 Number 6Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

Interestingly, spirochetes carrying bbb22 alone driven by the long5=UTR sequence (22distp-bbb22�) demonstrated statistically sig-nificantly increased levels of hypoxanthine uptake relative to thelevels of both the �bbb22-23 mutant and heat-killed spirochetes attime points 5, 8, 15, and 30 min, further suggesting a role for thelong 5=UTR sequence in bbb22 function. However, the maximumamount of hypoxanthine uptake achieved by 22distp-bbb22� spi-rochetes was significantly reduced compared to that of spirochetescarrying both the bbb22 and bbb23 genes (Fig. 3A). In addition,hypoxanthine uptake by 22distp-bbb22� spirochetes reached sat-uration in the first 5 min of the assay. In contrast, bbb22-bbb23�

spirochetes demonstrated an increase in hypoxanthine uptake outto 15 min prior to reaching saturation. These data suggest thatalthough spirochetes carrying bbb22 alone are capable of trans-porting hypoxanthine, this function is significantly reduced in theabsence of bbb23. We further examined the hypoxanthine trans-port function conferred by bbb22 alone by calculating the Km val-ues for hypoxanthine for clones 22distp-bbb22� and bbb22-bbb23� based on the initial rates of hypoxanthine uptake for theseclones across increasing concentrations of hypoxanthine. Wefound that the Km for hypoxanthine for 22distp-bbb22� spiro-chetes was no different from the Km for hypoxanthine for bbb22-bbb23� spirochetes (2.6 0.1 �M and 3.4 0.5 �M, respectively)(Table 3). These findings indicate that the absence of bbb23 in22distp-bbb22� spirochetes did not result in a significant altera-tion in the affinity for hypoxanthine uptake. As expected, the�bbb22-23 mutant and heat-killed spirochetes demonstrated noincrease in the amount of [3H]hypoxanthine uptake over the 30-min time course (Fig. 3A) (7). Similarly, no increase in theamount of [3H]hypoxanthine uptake was detected for the 23p-bbb23� and 22proxp-bbb22� spirochetes over the time period ofthe assay, suggesting that these clones were unable to transporthypoxanthine. The amounts of hypoxanthine transported by spi-rochetes carrying the bbb23 gene alone under the control of theconstitutive flaB promoter over the 30-min time course were notfound to be statistically different from the background levels de-tected in the �bbb22-23 mutant and heat-killed spirochetes. How-ever, unlike the �bbb22-23 mutant and heat-killed spirochetes,the flaBp-bbb23� spirochetes demonstrated a statistically signifi-cant increase in hypoxanthine uptake between the 10-s and 15-min time points (P � 0.0009, unpaired t test) (Fig. 3A), suggestingthat bbb23 alone may confer weak hypoxanthine transport ac-tivity. The initial rates of hypoxanthine uptake by flaBp-bbb23�

spirochetes were found to be too low to allow accurate deter-mination of the Km for hypoxanthine for this clone (data notshown). Together these data indicate that individually thebbb22 gene and possibly the bbb23 gene are each capable ofproviding B. burgdorferi with a minimal capacity for hypoxan-thine transport, suggesting a synergistic role for the two genestogether for the ability of B. burgdorferi to transport wild-typelevels of hypoxanthine.

The bbb22 gene alone allows spirochetes to take up wild-typesteady-state levels of adenine. It has been previously demon-strated that genes bbb22-23 contribute to the ability of B. burgdor-feri to undergo rapid transport of adenine (7) (Fig. 3B). In addi-tion, B. burgdorferi is able to transport adenine in the absence ofbbb22-23, albeit at a much lower rate, likely due to the function ofan additional, as of yet unknown, adenine transporter(s) (7)(Fig. 3B). To understand the individual contributions of bbb22and bbb23 to adenine uptake, the B. burgdorferi clones carryingonly one or the other gene were analyzed over a 30-min timecourse at a physiologically relevant concentration of [3H]adenineof 0.396 �M (22). Similar to what was observed for hypoxanthinetransport, clones 23p-bbb23� and 22proxp-bbb22� demonstratedno detectable increase in adenine transport relative to the�bbb22-23 mutant. The amounts of adenine uptake detected inthe two clones at the 30-min time point were significantly reducedcompared to the level in the �bbb22-23 mutant but not statisti-cally different from that in heat-killed spirochetes. The addition ofthe bbb23 gene under the control of the flaB promoter did notprovide the �bbb22-23 mutant with any increased ability to trans-port adenine. Together, these data indicate that bbb23 alone andbbb22 under the control of the proximal promoter do not conferadenine transport activity (Fig. 3B). In contrast, 22distp-bbb22�

spirochetes demonstrated a significant increase in adenine uptakecompared to that of the bbb22-23 mutant (Fig. 3B). Furthermore,with the exception of the 2-min time point, there was no statisticaldifference between the adenine uptake by 22distp-bbb22� spiro-chetes and that of spirochetes harboring both genes (Fig. 3B). The22distp-bbb22� spirochetes reached the same level of adenine sat-uration within the same amount of time as the bbb22-23� spiro-chetes. Kinetic analyses of the initial rates of adenine transport for22distp-bbb22� and bbb22-23� spirochetes determined that bothclones demonstrated the same Km values for adenine (0.19 �M 0.01 and 0.13 �M 0.04, respectively) (Table 3). Together, thesedata suggest that bbb23 alone does not contribute to adeninetransport under the conditions of this assay. The bbb22 gene, whenexpressed under the control of the long 5= UTR, was sufficient torestore adenine uptake to the level of bbb22-23� spirochetes.

Cold hypoxanthine, adenine, and guanine compete with[3H]hypoxanthine for uptake by spirochetes harboring eitherbbb22 or bbb23 alone. Taking advantage of the finding that spi-rochetes harboring bbb22 alone (22distp-bbb22�) or bbb23 alone(flaBp-bbb23�) were capable of some degree of [3H]hypoxanthineuptake (Fig. 3B), we sought to determine the ability of unlabeledhypoxanthine, adenine, and guanine to compete with this func-tion. Consistent with our data demonstrating that bbb22 is func-tional to transport adenine as well as hypoxanthine, [3H]hypox-anthine transport by the 22distp-bbb22� clone was reduced tonearly 3% and 20% in the presence of excess cold adenine and coldhypoxanthine, respectively (Fig. 4). Furthermore, bbb22 providedthe spirochetes with the ability to transport guanine as [3H]hy-poxanthine transport by the 22distp-bbb22� clone was reduced to37% in the presence of excess cold guanine. Similarly, flaBp-bbb23� spirochetes demonstrated a reduction in [3H]hypoxan-thine transport in the presence of excess cold hypoxanthine, ade-nine, and guanine (13%, 16% and 5%, respectively) (Fig. 4). Thefinding that that excess cold adenine was able to compete for[3H]hypoxanthine transport by flaBp-bbb23� spirochetes was sur-prising, given that direct transport of [3H]adenine by this clonewas not detected (Fig. 3B), and suggested that the low adenine

TABLE 3 Affinities for hypoxanthine and adenine of spirochetescarrying bbb22 alone or both bbb22 and bbb23

Clone

Km (�M)

Hypoxanthine Adenine

bbb22-23� 3.4 0.5 0.13 0.0422distp-bbb22� 2.6 0.10 0.19 0.01

BBB22 and BBB23 Transport Function

June 2015 Volume 83 Number 6 iai.asm.org 2229Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

transport activity conferred by bbb23 alone was detectable onlywhen the substrate was present in high concentrations. Alterna-tively, it is possible that adenine is not actively transported by theflaBp-bbb23� spirochetes and that high concentrations of adeninemay interfere with the hypoxanthine transport activity of thisclone. As expected based on our previous work (7), hypoxanthinetransport by either of these clones was unaffected by the presenceof the pyrimidine cytosine (Fig. 4). Overall, these data suggest thatindividually both bbb22 and bbb23 are able to provide B. burgdor-feri with the ability transport to hypoxanthine, adenine, and gua-nine. Moreover, these data suggest a trend in which adenine is astronger competitor than guanine for hypoxanthine transport bybbb22� spirochetes (P � 0.06, unpaired t test) and that guanine isa stronger competitor than adenine for hypoxanthine transport bybbb23� spirochetes (P � 0.02 unpaired t test).

bbb22, but not bbb23, is critical for mouse infectivity. To-gether, the bbb22 and bbb23 genes are essential for B. burgdorferiinfection in mice (7). We have now demonstrated that individu-ally bbb22 and bbb23 vary in their contributions to the ability of B.burgdorferi to transport purines (Fig. 3 and 4). To investigate theindividual contributions of bbb22 and bbb23 to mouse infection,mice were needle inoculated with B. burgdorferi clones bbb22-23�,�bbb22-23/vector, 23p-bbb23�, flaBp-bbb23�, 22proxp-bbb22�,and 22distp-bbb22� at a dose of 104 or 107 spirochetes. Mouseinfection was assessed 3 weeks postinoculation by serology andspirochete reisolation from tissues. Of the mice inoculated with B.burgdorferi clones carrying bbb23 or bbb22 alone, only the miceinoculated with 22distp-bbb22� spirochetes acquired infections(Table 3). These data suggest that bbb22 is sufficient for B. burg-dorferi infection of mice whereas bbb23 is not. In support of thisnotion, mice infected with spirochetes harboring bbb23 alone(flaBp-bbb23� and 23p-bbb23�) were found to be seronegativeand negative for reisolation of spirochetes from tissues (Table 3),

serology and reisolation of spirochetes from mouse tissues beingqualitative measures of B. burgdorferi infectivity. To identifyquantitative differences between the B. burgdorferi infections by22distp-bbb22� and bbb22-23� spirochetes, the spirochete loadsin mouse tissues inoculated with the two different clones wereevaluated by quantitative PCR. Our results demonstrated that spi-rochete loads in the ear tissues of mice infected with the 22distp-bbb22� clone were significantly reduced compared to the spiro-chete burdens in the ear tissues of mice infected with bbb22-23�

spirochetes (Fig. 5). Although no statistical difference was de-tected between the spirochete loads in the heart and joint tissuesinfected with the two clones, the data suggested a trend towardreduced spirochete numbers in the tissues of the mice infectedwith spirochetes harboring bbb22 alone (Fig. 5C). This analysiswas also carried out for tissues harvested from mice inoculatedwith clones flaBp-bbb23� and �bbb22-23, both of which werefound to be noninfectious by the qualitative measures of infection(Table 4). As expected, there was no quantitative detection of spi-rochetes in the tissues of mice inoculated with clones flaBp-bbb23� and �bbb22-23 (data not shown). Together, these datasuggest that spirochetes harboring bbb22 alone are able to infectmice. However, the bbb22 gene alone is not sufficient to allow thespirochetes to maintain the level of B. burgdorferi burden in tissuesachieved by spirochetes that carry both the bbb22 and bbb23 genestogether.

DISCUSSION

Borrelia burgdorferi is dependent upon purine salvage from theenvironment for the synthesis of DNA and RNA (2, 5, 6, 23). Wehave previously shown that bbb22 and bbb23 encode purine per-meases, which together are essential for B. burgdorferi infection inmice (7). In this study, we now demonstrate the individual con-tributions of bbb22 and bbb23 to purine transport and B. burgdor-feri mouse infectivity. To understand the individual roles of bbb22and bbb23, we first undertook transcriptional analysis to identifythe transcription start sites for both genes. 5= RACE analysis iden-tified a distinct transcription start site for bbb22. Surprisingly,however, reverse transcriptase PCR analysis across the intergenicregion between the annotated bbb23 and bbb22 ORFs indicatedthe presence of a continuous transcript across the two genes (S.Jain and M. W. Jewett, unpublished data). Together, these data

FIG 4 Adenine and guanine compete with hypoxanthine for transport byboth BBB23 and BBB22 individually. [3H]hypoxanthine (0.25 �M, 20�Ci)uptake by 22distp-bbb22� and flaBp-bbb23� spirochetes was determined at 15min following the addition of radiolabeled purine in the absence (no compet-itor) or presence of 20-fold excess (5 �M) unlabeled competitor as indicated.[3H]hypoxanthine uptake by 22distp-bbb22� or flaBp-bbb23� spirochetes inthe absence of competitor was taken as 100%. All other data are represented asthe percent uptake relative to that of 22distp-bbb22� or flaBp-bbb23� spiro-chetes, respectively, in the absence of competitor. Heat-killed spirochetesserved as the negative controls. Data sets were compared by an unpaired t testusing GraphPad Prism, version 5.0, and relevant P values are shown.

FIG 5 The bbb22 gene alone is not sufficient to maintain wild-type levels ofspirochete loads in infected mouse tissues. DNA was isolated from ear, heart,and joint tissues of C3H/HeN mice inoculated with 1 � 104 bbb22-23� or22distp-bbb22� spirochetes. Samples were analyzed for B. burgdorferi (Bb) flaBand murine nidogen DNA copies by qPCR. Each data point represents theaverage of triplicate measures from the DNA of an individual mouse. Themean value for each group is indicated by a horizontal line. Data sets werecompared by one-way ANOVA using GraphPad Prism, version 5.0. *, P 0.05; NS, not significant.

Jain et al.

2230 iai.asm.org June 2015 Volume 83 Number 6Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

suggested the possibility of an additional long bbb22 transcriptthat originates from either cotranscription with bbb23 or a secondbbb22 transcription start site within the bbb23 ORF. Indeed, the 5=end of a transcript was identified upstream of bbb22 within bbb23,just downstream of a promoter sequence identified in our B. burg-dorferi IVET screen (21; M. W. Jewett, unpublished data), indicat-ing that this transcript is expressed both in vitro and in vivo. Thelong 5= UTR sequence produced from the bbb22 distal promoter(22distp) was found to be important for bbb22 expression as well asbbb22-dependent purine transport and murine infectivity. A long5=UTR was also identified upstream of the bbb23 ORF. However,the bbb22-23 trans-complementation construct carrying only 110bp of bbb23 upstream sequence was sufficient to restore wild-typein vitro purine transport and mouse infectivity to �bbb22-23 spi-rochetes under the experimental conditions tested, raising thequestion as to the importance of the bbb23 5=UTR (7; this work).Regulation of genes that transport and metabolize purines hasbeen shown to occur by riboswitch-mediated control (24, 25). Forpurine-sensing riboswitches, this type of regulation typically oc-curs through the binding of guanine or adenine to the riboswitchsequence in the long 5=UTR sequences of the regulated genes (26).Currently, there are four known classes of purine-sensing ribo-switches; however, additional classes may remain yet undiscov-ered (26). The extended leader sequences detected upstream ofbbb22 and bbb23 suggest the possibility of a riboswitch mech-anism, but bioinformatics analysis (27) of the bbb22 and bbb235= UTR sequences did not identify these to be one of the fourknown classes of purine-sensing riboswitches. The functionalroles of the 5= UTR sequences of both bbb22 and bbb23 areunder investigation.

The transcription level of the bbb23 gene in spirochetes grownto late stationary phase was found to be approximately 100-foldless than that of the bbb22 gene. This disparity in the level ofexpression of the two genes in late stationary phase was eliminatedwhen the bbb23 ORF was expressed under the control of the con-stitutive flaB promoter. Although overexpression of bbb23 maynot reflect the physiological level of bbb23 expression, geneticstandardization of the bbb23 and bbb22 expression levels alloweddirect comparison of in vitro transport function(s) conferred byeach gene alone.

The bbb22 gene, only when expressed under the control of thedistal promoter within the bbb23 ORF (22distp-bbb22�), enabled�bbb22-23 spirochetes to take up [3H]hypoxanthine and [3H]ad-enine. Furthermore, [3H]hypoxanthine uptake by 22distp-bbb22�

spirochetes was inhibited by excess cold hypoxanthine and ade-nine as well as guanine, albeit to a lesser extent. The constitutiveexpression of the bbb23 gene (flaBp-bbb23�) enabled �bbb22-23spirochetes to take up a measurable amount of [3H]hypoxanthinebut not [3H]adenine. Surprisingly, however, [3H]hypoxanthineuptake by flaBp-bbb23� spirochetes was inhibited by excess coldhypoxanthine, adenine, and guanine. These findings suggest thatbbb23 confers purine transport function; however, detection ofthis function in vitro was achieved only under experimental con-ditions where the bbb23 gene was constitutively expressed andwhere the amount of adenine was greater than 10 times the phys-iological concentration (22). These data suggest that individuallyboth bbb22 and bbb23 are able to provide �bbb22-23 spirocheteswith the ability to transport hypoxanthine, adenine, and guanine,but the purine transport functions conferred by each gene arequantitatively distinct. Kinetic analysis revealed that there was nodifference between the Km values for hypoxanthine and adeninefor 22distp-bbb22� spirochetes and spirochetes carrying bothbbb22 and bbb23. In contrast, the Km values for hypoxanthine andadenine for spirochetes expressing only bbb23 (flaBp-bbb23�)were unable to be measured due to the low initial rates of hypo-xanthine and adenine uptake by this clone. These data suggest thatthe BBB23 protein does not contribute to the affinity of theBBB23-BBB22 purine transport system for hypoxanthine and ad-enine. Rather, results indicate a role of bbb23 in the ability ofbbb22-bbb23� spirochetes to achieve the maximum level of hypo-xanthine saturation. Because spirochetes carrying bbb22 alonedemonstrated no defect in the Km value for hypoxanthine yet thelevel of hypoxanthine saturation for this clone was significantlyreduced relative to that of bbb22-bbb23� spirochetes, the affinityof BBB22 for hypoxanthine may be sufficient for the initial uptakeof this purine, but continuous hypoxanthine uptake requires theadditional activity of the lower-affinity BBB23 protein. Together,these findings demonstrate that the purine transport function en-coded by the bbb22 gene alone is greater than the purine transportfunction encoded by the bbb23 gene alone. The bbb22 gene was

TABLE 4 Spirochetes carrying bbb22 alone demonstrate qualitative measures of mouse infectivity

CloneSpirochete doseper mouse

Serology (no. of seropositivemice/no. of mice inoculated)a

Spirochete reisolation from tissues(no. of positive tissues/no. of mice inoculated)

Ear Bladder Joint

bbb22-23� 1 � 104 6/6 6/6 6/6 6/61 � 107 6/6 6/6 6/6 6/6

�bbb22-23 1 � 104 0/6 0/6 0/6 0/61 � 107 0/6 0/6 0/6 0/6

22proxp-bbb22� 1 � 104 0/6 0/6 0/6 0/61 � 107 0/6 0/6 0/6 0/6

22distp-bbb22� 1 � 104 6/6 6/6 6/6 6/61 � 107 6/6 6/6 6/6 6/6

23p-bbb23� 1 � 104 0/6 0/6 0/6 0/61 � 107 0/6 0/6 0/6 0/6

flaBp-bbb23� 1 � 104 0/6 0/6 0/6 0/61 � 107 0/6 0/6 0/6 0/6

a Determined at 3 weeks postinoculation by positive serology against a total protein lysate of infectious B. burgdorferi strain B31 (1 � 104 spirochetes) or purified recombinantglutathione S-transferase (rGST)-P39 protein (1 � 107 spirochetes) (32).

BBB22 and BBB23 Transport Function

June 2015 Volume 83 Number 6 iai.asm.org 2231Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

sufficient to achieve wild-type levels of adenine uptake. However,neither gene alone conferred the level of hypoxanthine uptakeaccomplished by bbb22-23� spirochetes, demonstrating that bothgenes are required to provide wild-type purine transport function.

The amino acid sequences of BBB22 and BBB23 are 78.3%identical. Nonetheless, the two proteins demonstrate distinct abil-ities to transport hypoxanthine, adenine, and guanine. This sug-gests that the differential transport functions of the two proteinsmay be dictated by a few key differences in amino acid residues.Recently, structure-function analyses of the E. coli adenine-spe-cific permeases PurP and YicO and hypoxanthine-guanine-spe-cific permeases YjcD and YgfQ as well as the A. nidulans adenine-hypoxanthine-guanine permease AzgA identified key amino acidresidues critical for purine uptake activity (13, 15). Both of thesestudies identified an aspartic acid and a glutamic acid residue(Asp339 and Glu394 in AzgA) to be essential for the transportfunction for all five COG2252 family members (13, 15). Theseresidues are conserved in BBB22 and BBB23. Strikingly, however,two key residues shown to be important for purine transport func-tion are conserved in BBB22 but divergent in BBB23 (13, 15).Amino acid Thr275 in YjcD was found to be critical for hypoxan-thine-guanine transport (13, 15). This residue is conserved inBBB22 as well as in the other COG2252 proteins (13, 15), but anisoleucine is present at this position in BBB23. In addition, Gly129has been shown to be important for AzgA substrate binding(13, 15). This residue is conserved in BBB22 and in the E. coliadenine-specific permeases, whereas BBB23 and the E. coli hypox-anthine-guanine-specific permeases have a serine at this position(13, 15). Future mutational analysis of the BBB22 and BBB23proteins, guided by the current understanding of the molecularinteractions important for the transport activities of COG2252family members, will provide insight into the molecular basisfor the functional differences between these two highly relatedtransporters.

Antibodies specific for the BBB22 and BBB23 proteins are notcurrently available. Numerous attempts to detect BBB22 andBBB23 protein production in B. burgdorferi carrying epitope-tagged bbb22 and bbb23 under the control of both native andconstitutive promoters were unsuccessful (S. Jain and M. W. Jew-ett, unpublished data), suggesting that the amounts of BBB22 andBBB23 produced in in vitro-grown B. burgdorferi are below thelevels of detection of our assay. Nonetheless, low levels of BBB22and BBB23 protein have been detected previously through ge-nome-wide proteomic analysis of in vitro-grown B. burgdorferiusing nano-liquid chromatography and tandem mass spectrome-try (28). Future studies will focus on sensitive detection of BBB22and BBB23 using unique monoclonal antibodies to experimen-tally establish the cellular localization of the two proteins and de-termine whether the purine transport activities of these proteinsinvolve interaction with one another and/or other yet to be iden-tified B. burgdorferi proteins.

Of the B. burgdorferi clones carrying only bbb22 or bbb23, only22distp-bbb22� spirochetes demonstrated mouse infectivity asmeasured by serology and reisolation of spirochetes from infectedtissues. This finding highlights the requirement for the bbb22 5=UTR sequence for bbb22 function and the distinct individual ca-pabilities of the bbb22 and bbb23 genes. Both hypoxanthine andadenine are available for salvage from the blood and tissues (22).B. burgdorferi is transiently present in the blood during the initialstages of mammalian infection and then rapidly disseminates to

distal tissues (29–31). Hypoxanthine has been reported to be themost abundant purine in human plasma, at a concentration ofapproximately 8 �M (20). The concentration of adenine in hu-man plasma has been found to be approximately 0.3 �M (20). Theconcentration of both hypoxanthine and adenine in tissues hasbeen reported to be approximately 0.4 �M (22). We have foundthat the Km values for hypoxanthine and adenine for spirochetescarrying both the bbb22 and bbb23 genes are 3.4 0.5 �M and0.13 0.04 �M, respectively. These Km values correlate well withthe reported physiological concentrations of hypoxanthine inplasma and of adenine in plasma and tissues. Because the reportedphysiological concentration of hypoxanthine in tissues is 10-foldless than the Km, this suggests that following the transient blood-borne phase of dissemination of B. burgdorferi may be primarilydependent on adenine salvage to generate hypoxanthine throughadenine deamination (5, 8, 9). Indeed, our findings indicate thatbbb22-dependent adenine transport is essential for B. burgdorferisurvival in mice. The affinity of 22distp-bbb22� spirochetes forhypoxanthine and adenine was equivalent to that of spirochetescarrying both genes. However, the maximum amount of hypoxan-thine uptake achieved by 22distp-bbb22� spirochetes was significantlyreduced compared to the amount with the bbb22-bbb23� clone. Theenzymatic conversion of adenine to hypoxanthine by adeninedeaminase, AdeC, offers B. burgdorferi another source of hypo-xanthine (8). Nevertheless, our data suggest that this deficiency inhypoxanthine uptake resulted in a quantitative difference in theinfectious phenotype of 22distp-bbb22� versus that of bbb22-23�

spirochetes. Tissues isolated from mice infected with 22distp-bbb22� spirochetes demonstrated a reduced spirochete load com-pared to tissues isolated from mice infected with bbb22-23� spi-rochetes, perhaps suggesting that the transport of maximumamounts of hypoxanthine during the initial stages of infection isimportant for B. burgdorferi dissemination and colonization oftissues. Although flaBp-bbb23� spirochetes showed some abilityto transport purines in vitro, the finding that flaBp-bbb23� spiro-chetes are noninfectious in mice demonstrated that the purinetransport affinities of spirochetes containing only bbb23 are notsufficient to support spirochete survival. The bbb23 gene is thusdispensable for B. burgdorferi survival in mice. Nonetheless, thisgene is required along with bbb22 for B. burgdorferi to achievemaximal spirochete loads in infected mouse tissues.

The combined significance of the bbb22 and bbb23 genes to B.burgdorferi biology is highlighted by the fact that these genes areencoded on the essential cp26 plasmid. cp26, unlike the approxi-mately 20 linear and circular plasmids of B. burgdorferi, is presentin all natural isolates examined (23), indicating that the bbb22 andbbb23 genes are maintained across the B. burgdorferi population.All in all, these data indicate that both bbb22 and bbb23 are criticalfor B. burgdorferi to achieve maximal infection of mice and suggestthat the differences in the patterns of expression and functionalcapabilities of the highly related transporters may reflect distinctpurine salvage needs that the spirochete encounters throughoutits natural infectious cycle.

ACKNOWLEDGMENTS

We thank members of the Jewett lab for critical reading of the manuscript,Travis Jewett and Victor Davidson for insightful comments and sugges-tions, and Philip Adams for technical assistance. We also thank the NonaAnimal Facility animal care staff.

Research reported in this publication was supported by the National

Jain et al.

2232 iai.asm.org June 2015 Volume 83 Number 6Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

Institute of Allergy and Infectious Diseases of the National Institutes ofHealth under award numbers K22AI081730 (M.W.J.) and R01AI099094(M.W.J.), the National Research Fund for Tick-Borne Diseases (M.W.J.),and a UCF ORC in-house research grant (M.W.J.).

REFERENCES1. Kose M, Schiedel AC. 2009. Nucleoside/nucleobase transporters: drug

targets of the future? Future Med Chem 1:303–326. http://dx.doi.org/10.4155/fmc.09.29.

2. Barbour AG, Putteet-Driver AD, Bunikis J. 2005. Horizontally acquiredgenes for purine salvage in Borrelia spp. causing relapsing fever. Infect Im-mun 73:6165–6168. http://dx.doi.org/10.1128/IAI.73.9.6165-6168.2005.

3. Casjens S, Palmer N, van Vugt R, Huang WM, Stevenson B, Rosa P,Lathigra R, Sutton G, Peterson J, Dodson RJ, Haft D, Hickey E, GwinnM, White O, Fraser CM. 2000. A bacterial genome in flux: the twelvelinear and nine circular extrachromosomal DNAs in an infectious isolateof the Lyme disease spirochete Borrelia burgdorferi. Mol Microbiol 35:490 –516. http://dx.doi.org/10.1046/j.1365-2958.2000.01698.x.

4. Fraser CM, Casjens S, Huang WM, Sutton GG, Clayton R, Lathigra R,White O, Ketchum KA, Dodson R, Hickey EK, Gwinn M, Dougherty B,Tomb J F, Fleischmann RD, Richardson D, Peterson J, Kerlavage AR,Quackenbush J, Salzberg S, Hanson M, van Vugt R, Palmer N, AdamsMD, Gocayne J, Weidman J, Utterback T, Watthey L, McDonald L,Artiach P, Bowman C, Garland S, Fuji C, Cotton MD, Horst K, RobertsK, Hatch B, Smith HO, Venter JC. 1997. Genomic sequence of a Lymedisease spirochaete, Borrelia burgdorferi. Nature 390:580 –586. http://dx.doi.org/10.1038/37551.

5. Lawrence KA, Jewett MW, Rosa PA, Gherardini FC. 2009. Borreliaburgdorferi bb0426 encodes a 2=-deoxyribosyltransferase that plays a cen-tral role in purine salvage. Mol Microbiol 72:1517–1529. http://dx.doi.org/10.1111/j.1365-2958.2009.06740.x.

6. Pettersson J, Schrumpf ME, Raffel SJ, Porcella SF, Guyard C, LawrenceK, Gherardini FC, Schwan TG. 2007. Purine salvage pathways amongBorrelia species. Infect Immun 75:3877–3884. http://dx.doi.org/10.1128/IAI.00199-07.

7. Jain S, Sutchu S, Rosa PA, Byram R, Jewett MW. 2012. Borrelia burg-dorferi harbors a transport system essential for purine salvage and mam-malian infection. Infect Immun 80:3086 –3093. http://dx.doi.org/10.1128/IAI.00514-12.

8. Jewett MW, Lawrence K, Bestor AC, Tilly K, Grimm D, Shaw P,VanRaden M, Gherardini F, Rosa PA. 2007. The critical role of the linearplasmid lp36 in the infectious cycle of Borrelia burgdorferi. Mol Microbiol64:1358 –1374. http://dx.doi.org/10.1111/j.1365-2958.2007.05746.x.

9. Jewett MW, Lawrence KA, Bestor A, Byram R, Gherardini F, Rosa PA.2009. GuaA and GuaB are essential for Borrelia burgdorferi survival in thetick-mouse infection cycle. J Bacteriol 191:6231– 6241. http://dx.doi.org/10.1128/JB.00450-09.

10. Cecchetto G, Amillis S, Diallinas G, Scazzocchio C, Drevet C. 2004. TheAzgA purine transporter of Aspergillus nidulans. Characterization of a pro-tein belonging to a new phylogenetic cluster. J Biol Chem 279:3132–3141.http://dx.doi.org/10.1074/jbc.M308826200.

11. Goudela S, Tsilivi H, Diallinas G. 2006. Comparative kinetic analysis ofAzgA and Fcy21p, prototypes of the two major fungal hypoxanthine-adenine-guanine transporter families. Mol Membr Biol 23:291–303. http://dx.doi.org/10.1080/09687860600685109.

12. Kozmin SG, Stepchenkova EI, Chow SC, Schaaper RM. 2013. A criticalrole for the putative NCS2 nucleobase permease YjcD in the sensitivity ofEscherichia coli to cytotoxic and mutagenic purine analogs. mBio4:e00661-13. http://dx.doi.org/10.1128/mBio.00661-13.

13. Krypotou E, Diallinas G. 2014. Transport assays in filamentous fungi:kinetic characterization of the UapC purine transporter of Aspergillus ni-dulans. Fungal Genet Biol 63:1– 8. http://dx.doi.org/10.1016/j.fgb.2013.12.004.

14. Mansfield TA, Schultes NP, Mourad GS. 2009. AtAzg1 and AtAzg2comprise a novel family of purine transporters in Arabidopsis. FEBS Lett583:481– 486. http://dx.doi.org/10.1016/j.febslet.2008.12.048.

15. Papakostas K, Botou M, Frillingos S. 2013. Functional identification ofthe hypoxanthine/guanine transporters YjcD and YgfQ and the adenine

transporters PurP and YicO of Escherichia coli K-12. J Biol Chem 288:36827–36840. http://dx.doi.org/10.1074/jbc.M113.523340.

16. Saxild HH, Brunstedt K, Nielsen KI, Jarmer H, Nygaard P. 2001.Definition of the Bacillus subtilis PurR operator using genetic and bioin-formatic tools and expansion of the PurR regulon with glyA, guaC, pbuG,xpt-pbuX, yqhZ-folD, and pbuO. J Bacteriol 183:6175– 6183. http://dx.doi.org/10.1128/JB.183.21.6175-6183.2001.

17. Beaman TC, Hitchins AD, Ochi K, Vasantha N, Endo T, Freese E. 1983.Specificity and control of uptake of purines and other compounds in Ba-cillus subtilis. J Bacteriol 156:1107–1117.

18. Elias AF, Stewart PE, Grimm D, Caimano MJ, Eggers CH, Tilly K, BonoJL, Akins DR, Radolf JD, Schwan TG, Rosa P. 2002. Clonal polymor-phism of Borrelia burgdorferi strain B31 MI: implications for mutagenesisin an infectious strain background. Infect Immun 70:2139 –2150. http://dx.doi.org/10.1128/IAI.70.4.2139-2150.2002.

19. Eells JT, Spector R. 1983. Determination of ribonucleosides, deoxyribo-nucleosides, and purine and pyrimidine bases in adult rabbit cerebrospi-nal fluid and plasma. Neurochem Res 8:1307–1320. http://dx.doi.org/10.1007/BF00964000.

20. Hartwick RA, Assenza SP, Brown PR. 1979. Identification and quanti-tation of nucleosides, bases and other UV-absorbing compounds in se-rum, using reversed-phase high-performance liquid chromatography. I.Chromatographic methodology. J Chromatogr 186:647– 658.

21. Ellis TC, Jain S, Linowski AK, Rike K, Bestor A, Rosa PA, Halpern M,Kurhanewicz S, Jewett MW. 2013. In vivo expression technology identi-fies a novel virulence factor critical for Borrelia burgdorferi persistence inmice. PLoS Pathog 9:e1003567. http://dx.doi.org/10.1371/journal.ppat.1003567.

22. Traut TW. 1994. Physiological concentrations of purines and pyrimidines. MolCell Biochem 140:1–22. http://dx.doi.org/10.1007/BF00928361.

23. Jewett MW, Byram R, Bestor A, Tilly K, Lawrence K, Burtnick MN,Gherardini F, Rosa PA. 2007. Genetic basis for retention of a criticalvirulence plasmid of Borrelia burgdorferi. Mol Microbiol 66:975–990. http://dx.doi.org/10.1111/j.1365-2958.2007.05969.x.

24. Mandal M, Boese B, Barrick JE, Winkler WC, Breaker RR. 2003.Riboswitches control fundamental biochemical pathways in Bacillus sub-tilis and other bacteria. Cell 113:577–586. http://dx.doi.org/10.1016/S0092-8674(03)00391-X.

25. Mandal M, Breaker RR. 2004. Adenine riboswitches and gene activationby disruption of a transcription terminator. Nat Struct Mol Biol 11:29 –35.http://dx.doi.org/10.1038/nsmb710.

26. Kim JN, Breaker RR. 2008. Purine sensing by riboswitches. Biol Cell100:1–11. http://dx.doi.org/10.1042/BC20070088.

27. Chang TH, Huang HD, Wu LC, Yeh CT, Liu BJ, Horng JT. 2009.Computational identification of riboswitches based on RNA conservedfunctional sequences and conformations. RNA 15:1426 –1430. http://dx.doi.org/10.1261/rna.1623809.

28. Angel TE, Luft BJ, Yang X, Nicora CD, Camp DG, II, Jacobs JM, SmithRD. 2010. Proteome analysis of Borrelia burgdorferi response to environ-mental change. PLoS One 5:e13800. http://dx.doi.org/10.1371/journal.pone.0013800.

29. Dolan MC, Piesman J, Schneider BS, Schriefer M, Brandt K, ZeidnerNS. 2004. Comparison of disseminated and nondisseminated strains ofBorrelia burgdorferi sensu stricto in mice naturally infected by tick bite.Infect Immun 72:5262–5266. http://dx.doi.org/10.1128/IAI.72.9.5262-5266.2004.

30. Liveris D, Schwartz I, Bittker S, Cooper D, Iyer R, Cox ME, WormserGP. 2011. Improving the yield of blood cultures from patients with earlyLyme disease. J Clin Microbiol 49:2166 –2168. http://dx.doi.org/10.1128/JCM.00350-11.

31. Rego RO, Bestor A, Stefka J, Rosa PA. 2014. Population bottlenecksduring the infectious cycle of the Lyme disease spirochete Borrelia burg-dorferi. PLoS One 9:e101009. http://dx.doi.org/10.1371/journal.pone.0101009.

32. Halpern MD, Molins CR, Schriefer M, Jewett MW. 2014. Simple objec-tive detection of human Lyme disease infection using immuno-PCR and asingle recombinant hybrid antigen. Clin Vaccine Immunol 21:1094 –1105.http://dx.doi.org/10.1128/CVI.00245-14.

BBB22 and BBB23 Transport Function

June 2015 Volume 83 Number 6 iai.asm.org 2233Infection and Immunity

on January 18, 2020 by guesthttp://iai.asm

.org/D

ownloaded from