1

New recombinant BCG expression vectors: improving genetic control over 1

mycobacterial promoters 2

3

4

Alex I. Kannoab, Cibelly Goulartab, Henrique K. Rofattoab, Sergio C. Oliveirac, Luciana 5

C. C. Leitea, Johnjoe McFaddend# 6

7

Centro de Biotecnologia, Instituto Butantan, São Paulo, SP, Brazila; Programa de Pós 8

Graduação Interunidades em Biotecnologia USP – IPT – IB, São Paulo, Brazilb, 9

Departamento de Bioquímica e Imunologia, Universidade Federal de Minas Gerais, Belo 10

Horizonte, MG, Brazilc; Department of Microbial and Cellular Sciences, Faculty of 11

Health and Medical Sciences, University of Surrey, Guildford, Surrey, United Kingdomd. 12

13

Running Head: New recombinant BCG expression vectors 14

15

16

17

18

19

20

# Address correspondence to Johnjoe McFadden, j.mcfadden@surrey.ac.uk 21

AEM Accepted Manuscript Posted Online 5 February 2016Appl. Environ. Microbiol. doi:10.1128/AEM.03677-15Copyright © 2016, American Society for Microbiology. All Rights Reserved.

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

2

Abstract 22

23

The expression of many antigens, stimulatory molecules or even metabolic pathways in 24

mycobacteria such as Mycobacterium bovis BCG or M. smegmatis was made possible 25

through the development of shuttle-vectors and several recombinant vaccines have been 26

constructed. However, gene expression in any of these systems relied mostly on selection 27

of natural promoters expected to provide the required level of expression by trial and 28

error. To establish a systematic selection of promoters with a range of strengths, we 29

generated a library of mutagenized promoters through error-prone PCR of the strong PL5 30

promoter, originally from the mycobacteriophage L5. These promoters were cloned 31

upstream of the enhanced green fluorescent protein (eGFP) gene reporter and 32

recombinant M. smegmatis were identified exhibiting a wide range of fluorescence. A set 33

of promoters was selected and identified as high (pJK-F8), intermediate (pJK-B7, pJK-34

E6, pJK-D6) and low (pJK-C1) promoter strengths in both M. smegmatis and M. bovis 35

BCG. The sequencing of the promoter region demonstrated it was extensively modified 36

(6-11%) in all selected plasmids. To test the functionality of the system two different 37

expression vectors were demonstrated to allow corresponding expression levels of the 38

Schistosoma mansoni antigen, Sm29, in BCG. The approach conducted here can be 39

useful for adjusting expression levels for synthetic and/or systems biology studies or for 40

vaccine development to maximize the immune response. 41

42

Keywords: Recombinant BCG, promoters, random mutagenesis, vaccine. 43

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

3

Introduction 44

45

BCG is currently the world’s most widely used vaccine and has been given to 46

more than three billion people, making it a very attractive prospect for development of a 47

live recombinant BCG (rBCG) multivaccine (1). The first generation of rBCG vaccines 48

was developed in the 1990’s as rBCG strains that expressed homologous and 49

heterologous antigens from a wide range of pathogens (2-5). Stability of the heterologous 50

(or native) gene(s) in BCG was usually obtained by cloning it on a plasmid or 51

chromosomally integrative vector with expression achieved by placing it under the 52

control of a range of mycobacterial promoters (2). The heat shock protein promoters 53

(PHSP60 from BCG and PHSP70 from M. tuberculosis), the IS900 ORF promoter (PAN from 54

M. paratuberculosis) and the mutated β-lactamase promoter (PBlaF* from M. fortuitum) 55

are the most used ones, as reviewed by (2). However, the lack of knowledge on the 56

transcriptional mechanisms regulating mycobacterial promoters, results in unpredictable 57

gene expression levels. Yet level of antigen expression is likely to be a crucial factor in 58

both the strength and the pattern of subsequent immune responses. For example, a study 59

in which expression of the M. tuberculosis Antigen 85B (Ag85B) in rBCG was placed 60

under the control of a limited set of promoters, found that increasing promoter activity 61

caused a skewing of the immune response to Ag85B in mice from a mixed Th1/Th2 to a 62

predominantly Th1 response (6). In another study, increasing expression of M. 63

tuberculosis 19 kDa lipoprotein was found to lead to a complete abrogation of the 64

protective efficacy of BCG by polarizing the host immune responses to the Th2 subtype 65

(7). Another important factor for any vaccine vector is stability and the associated 66

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

4

metabolic burden, which can lead to loss of antigen expression and/or premature 67

elimination of the vector in the host due to loss of fitness (8). Indeed, loss of vector 68

during replication in the host is thought to have been responsible for the failure of a 69

rBCG vaccine expressing the outer surface protein A (OspA) of Borrelia burgdorferi to 70

generate an appropriate immune response in a clinical trial, despite the apparent efficacy 71

of the vaccine in mice (9). Our own earlier studies similarly demonstrated vector 72

instability and a lack of correlation of immune response in mice with expression levels in 73

vitro (10). 74

To improve the genetic control on gene expression in mycobacteria, different 75

tools and strategies can be used, such as inducible promoters (11), synthetic genes with 76

codon optimized sequences (12, 13), substitution of the initiation codon (14), addition of 77

Shine-Dalgarno sequences (15), increase in plasmid copy number (16) or integration into 78

the mycobacterial genome (17). However, once again, expression levels are often 79

unpredictable and may be further complicated by regulatory influences on expression 80

levels, particularly when growing in the host environment. In this work, as a first step 81

towards a more rational approach to vaccine vector design we engineered the 82

mycobacteriophage promoter PL5 (18). Our aim was to generate a set of mycobacteria 83

promoters that can be used to obtain a predictable range of gene expression levels in both 84

M. smegmatis and M. bovis BCG. 85

86

Materials and Methods 87

88

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

5

Ethics Statement. This study was conducted in accordance with the recommendations and 89

approval by the Committee of Ethical in Animal Experiments of the Butantan Institute 90

under the protocol 594/09. 91

Bacterial strains and plasmids. E. coli DH5α (19) (Invitrogen) was used in all cloning 92

steps using lysogeny broth (LB) or LB plates with kanamycin (20 µg/mL) for selection of 93

transformants. M. smegmatis MC2 155 strain (20) and M. bovis BCG strain Pasteur (21) 94

were grown at 370C in Difco Middlebrook 7H9 broth (Becton, Dickinson) enriched with 95

10% (vol/vol) oleic acid-albumin-dextrose-catalase (OADC), 0.5% glycerol and 0.05% 96

Tween 80 (MB7H9) containing kanamycin when necessary. Eletrocompetent M. smegmatis 97

and BCG were transformed by electroporation as previously described (22) and 98

transformants selected in Middlebrook 7H10 agar plates with OADC (MB7H10) and 99

kanamycin. All primers were purchased from Eurofins MWG Operon (Ebersberg, 100

Germany) and all restriction enzymes from New England Biolabs (Hitchin, UK). GoTaq 101

DNA polymerase (Promega, Southampton, UK) was used in all PCR reactions. The 102

expression cassette of the pBRL8 plasmid (kindly provided by Dr. William Jacobs Jr., 103

Yeshiva University, NY) containing the PL5 promoter and egfp gene was PCR amplified 104

using primers 5’-TAGGGTACCTCTAGAGGAAACAGCTATGACCAT-3’ (Kpn I and 105

Xba I restriction sites underlined and in bold) and 5’-106

TAGATCGATCAGCTGTTACTTGTACAGCTCGT-3’ (Cla I and Pvu II restriction sites 107

underlined and in bold). This fragment was cloned into pJH152 (kindly provided by Dr. 108

Stewart T. Cole, Ecolé Polytechnique Fédérale de Lausanne, France) between restriction 109

sites Kpn I and Pvu II, thus generating pJH-egfp (Figure 1A). Next, error-prone PCR of the 110

promoter sequence was carried out in the presence of 25 µM of 8-oxo-dGTP and 5 µM of 111

dPTP (Jena Bioscience, Germany), along with 400 µM of natural dNTPs (Figure 1B). 112

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

6

According to manufacturer’s instructions, 20 rounds of amplification were performed using 113

primers 5’-TAGGTTTAAACAAACGGAAACAGCTATGACCAT-3’ (Pme I restriction 114

site in bold) and 5’-TAGCATATGCGATCTCCCTTTCCCGT-3’ (Nde I restriction site in 115

bold). Plasmid pJH-egfp and the PCR product were digested with Pme I and Nde I, resolved 116

in an agarose gel electrophoresis to remove the original cassette and clean up the PCR 117

reaction, and further purified using QIAquick PCR Purification Kit (Qiagen, Hilden, 118

Germany). Before ligation, the digested vector was treated with CIAP (Promega), 119

according to the manufacturer instructions for the dephosphorylation of 5’ overhangs and 120

purified once again. Digested fragments were allowed to ligate using T4 DNA ligase 121

(Promega) at 16°C for 16 h and used to transform E. coli DH5α. Plasmid DNA from 122

approximately 2,000 colonies was extracted in pool generating the pJK plasmid library, 123

which in turn was used in the electroporation of M. smegmatis MC2 155 using a BioRad 124

Gene Pulser II apparatus (Hercules, USA). Transformants were allowed to grow for 3-5 125

days in MB7H10-kanamycin plates (Figure 1B). The plasmid pJH137 was digested using 126

Xba I and Nde I restriction enzymes to remove the heat shock promoter PHsp60 and insert in 127

place of PL5 at pJH-egfp generating pJHsp60. Plasmid pET-21 containing the Sm29 128

sequence to produce rSm29 protein was previously constructed (23). 129

Screening fluorescent M. smegmatis. After eye-visualization of M. smegmatis 130

transformants, approximately 200 colonies were isolated and grown in MB7H9-kanamycin. 131

Samples were allowed to grow for 16 h and screened in 96-well black plates in duplicates 132

using a Perkin Elmer Victor 3 fluorescence multiplate reader (Waltham, USA) setup to 485 133

nm/535 nm (excitation/emission). Also, the O.D.620 of each sample was measured. The 134

fluorescence of the strains was initially screened by fluorescence units (F.U.) and then, 135

following normalization by O.D., 5 strains were selected based on measurements of relative 136

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

7

fluorescence units (R.F.U.). The plasmid in each of these transformants was extracted using 137

illustra plasmidPrep Mini Spin kit (GE Healthcare Life Sciences, Freiburg, Germany) and 138

PL5 promoters sequenced using the primer 5’-TCAGCTTGCCGTAGGTGGCA-3’. 139

Characterization of GFP expression and in vitro growth of selected mutants. Shaking 140

flasks with MB7H9-kanamycin were inoculated with selected M. smegmatis::pJK in 141

triplicates to an O.D. of ~ 0.1. Culture turbidity and fluorescence were monitored every 1 h. 142

For comparison, the wild-type M. smegmatis mc2 155 strain grown in MB7H9 was also 143

analyzed. A FACSCanto II flow cytometer (Becton, Dickinson) was used to determine the 144

single-cell fluorescence of gated bacterial cells based on the median of FITC (using 488 145

excitation/500-560 emission filter). Turbidity was measured using a Ultrospec 2000 146

spectrophotometer (Pharmacia Biotech, Sweden). To compare the newly generated 147

promoters with a widely used one the PHsp60 was inserted in place of PL5. M. bovis BCG 148

containing these plasmids were analyzed and compared with M. smegmatis constructs. Late 149

log-phase cultures of recombinant M. smegmatis strains were disrupted on ice using a GE 150

100 Ultrasonic Processor at half-maximum constant output and protein concentration 151

determined in culture lysates by the DC Protein Assay (Biorad) using bovine serum 152

albumin as standard. Approximately 10 µg of soluble protein extracts were separated by 153

SDS-PAGE and stained with Coomassie Brilliant Blue (Biorad). The predicted molecular 154

weight of GFP is 28.3 kDa (http://web.expasy.org/compute_pi/). Following destaining the 155

band corresponding to GFP’s molecular weight was analyzed by densitometry using 156

ImageJ software. 157

Microscopy analysis of selected strains. M. smegmatis::pJK strains were observed under a 158

microscope. Late log-phase cultures were washed twice with PBS and placed on 159

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

8

microscope slides. Images were acquired on a LSM-510 META confocal system (Zeiss, 160

Germany) using 488 nm laser for excitation and BP500-550 filter for emission. 161

Fine-tuned expression of foreign antigen. To assess the usefulness of the pJK library 162

regardless of the downstream gene under its regulation, we used three plasmids containing 163

the mutagenized promoters, pJK-F8, pJK-B7 and pJK-C1. For this model, we removed the 164

egfp gene by digestion with Nde I and Pvu II and cloned the partial sequence of the 165

Schistosoma mansoni gene, sm29 (GenBank accession no. AF029222) codon-optimized for 166

expression in mycobacteria, as this antigen is currently considered a vaccine candidate (23), 167

creating pJK-F8.sm29, pJK-B7.sm29 and pJK-C1.sm29. The codon optimization was 168

performed by the DNA 2.0 company (CA, USA) according to their patented algorithm. 169

Codon-optimized sm29 had an increased GC content (from 37% to 51%). BCG was 170

transformed with these constructs and allowed to grow as described previously. 171

Kanamycin-resistant clones were grown until late-log phase, disrupted on ice and 172

quantified as previously described. Approximately 25 µg of total protein extracts were 173

separated by SDS-PAGE, proteins electrotransferred to a PVDF membrane (GE 174

Healthcare) and blocked for 16 h at 4oC in a 5% nonfat dry milk solution. Western blotting 175

analysis of Sm29 expression used polyclonal mouse anti-rSm29 antibody (1:1,000) and 176

horseradish peroxidase (HRP)-conjugated anti-mouse antibody adsorbed for human serum 177

proteins (KPL, Maryland, USA) (1:3,000). The chemiluminescent signal was detected 178

using the Immobilon Western Chemiluminescent HRP Substrate (Merck, Darmstadt, 179

Germany) and images acquired using the LAS4000 digital imaging system (GE). 180

Recombinant Sm29 protein and anti-rSm29 antibody production. Recombinant Sm29 181

was expressed in E. coli BL21 DE3 (Invitrogen) transformed with pET-21 containing the 182

Sm29 cDNA sequence and purified using metal affinity chromatography as previously 183

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

9

described (24). To generate the primary antibody, 3 BALB/c mice were immunized 184

subcutaneously with 3 doses of rSm29 (25 µg) adsorbed in aluminum hydroxide (250 µg) 185

with a 2 weeks interval between doses. Fifteen days after the last dose a cardiac puncture of 186

anesthetized mice was performed for blood collection and serum processing. 187

Statistical analysis. For statistical analysis, the software GraphPad Prism 5 (La Jolla, CA, 188

USA) was used. The relationship between pJH-egfp and pJK library in the initial screening 189

used Two-tailed Spearman’s Rank Correlation. The fluorescence induced by pJK in M. 190

smegmatis or BCG was analyzed by one-way ANOVA and Tukey’s post-test. Correlation 191

of interspecies fluorescence was tested using a non-parametric Spearman test. The growth 192

curves were compared using non-linear regression. 193

194

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

10

Results 195

196

Generation of the PL5 promoter library. First we used error-prone PCR (25) to generate a 197

promoter library of PL5 and assess variant sequences in their strengths by cloning the 198

promoters to drive expression of the enhanced green fluorescent protein, eGFP (Figure 1B). 199

The wild-type PL5 promoter and egfp gene were cloned into pJH152 generating pJH-egfp 200

(Figure 1A), which was used to transform M. smegmatis, thus representing the fluorescence 201

induced by the original promoter, which correlated with bacterial growth (Figure 2A). The 202

pJK plasmid library was used to transform M. smegmatis and the approximately 200 203

colonies named pJK-XA1 to -YH12 were screened by fluorimetry and demonstrated a wide 204

variation in fluorescence, ranging between 103 and 106 fluorescence units (F.U.) (Figure 205

2B). 206

Characterization of selected M. smegmatis::pJK. Based on the initial screening, 5 strains 207

were selected covering the fluorescence range for further analysis. First, the sequencing of 208

selected promoters revealed multiple mutations with modification in 6 to 11% of their 209

sequence (Figure 3). Since the O.D. of M. smegmatis::pJK library mutants varied between 210

0.3 to 1.5 the data was normalized and fluorescence plotted by relative units (R.F.U.) 211

(Figure 4A). Likewise, confocal microscopy confirmed the differential expression of GFP 212

where the fluorescence was distributed throughout the bacillus cell (Figure 4B). To exclude 213

the possibility that the differential fluorescence would be a consequence of impaired growth 214

or a distinct rate of GFP expression, selected M. smegmatis::pJK were evaluated for single-215

cell fluorescence during growth. Constant fluorescence is observed during growth in each 216

mutant strain (not shown). The M. smegmatis::pJK-C1 was the only one demonstrated to be 217

significantly different from the growth of the wild-type strain (Supplementary Figure 1). 218

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

11

Flow cytometry of M. smegmatis and M. bovis BCG::pJK strains. The selected pJK 219

plasmids were transformed into M. bovis BCG and late-log phase cultures analyzed 220

alongside recombinant M. smegmatis by flow cytometry. After gating the bacilli’s 221

population, fluorescence was determined by the median of FITC value. M. smegmatis 222

transformed with pJK-F8 showed high fluorescence; pJK-B7, pJK-E6 and pJK-D6 were 223

intermediate and pJK-C1 was strikingly low (Figure 5A). Recombinant M. smegmatis 224

strains were also submitted to SDS-PAGE and the bands corresponding to GFP were 225

quantified by densitometry (Figure 5B). The results demonstrated that the fluorescence 226

intensity measured by flow cytometry correlated with the amount of protein produced (r 227

value = 0.96) according to the strength of each promoter. BCG transformed with the same 228

plasmids revealed a high correlation with the fluorescence observed for M. smegmatis (r 229

value = 0.94) (not shown) where all mutants maintained their referred high, intermediate or 230

low strength in both species. pJHsp60 demonstrated a strength comparable to the lowest 231

expression vector pJK-C1 (Figure 6). 232

Mutant promoters induce differential antigen expression. To investigate the usefulness 233

of these promoters for the expression of other heterologous genes, we replaced the egfp 234

gene from pJK-F8, pJK-B7 and pJK-C1 with the codon optimized sm29 gene from S. 235

mansoni. Since, pJK-F8 produced no colonies after transformation, further work was 236

carried on using pJK-B7 and pJK-C1. Recombinant BCG clones of these strains were 237

grown to late-log phase and lysed by sonication. Total protein extracts were analyzed by 238

Western blot using specific anti-rSm29 antibodies. As predicted, we observed a more 239

intense immunoblot in the rBCG::pJK-B7.Sm29 extracts (higher promoter activity) as 240

compared to the reaction in the rBCG::pJK-C1.Sm29 extracts (lower promoter activity) at 241

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

12

the same molecular weight as the recombinant protein rSm29, but no immunoreaction is 242

observed in the wild type BCG protein extracts (Figure 7). 243

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

13

Discussion 244

245

Here we used the enhanced GFP (eGFP) as gene reporter to measure the differential 246

expression of this protein induced by mutations in the regulatory region of the 247

mycobacteriophage L5 PL5 promoter sequence. The PL5 promoter regulatory sequence was 248

subjected to random mutagenesis and a library of promoters exhibiting different strengths 249

measured by fluorescence was selected. The GFP has been one of the most used gene 250

reporters to characterize promoter strength among many organisms in a quantitative fashion 251

(26-28), including in mycobacterial species for promoter discovery, strength and trafficking 252

in the host (11, 29-32). 253

The observed 6-11% of mutations in the regulatory sequence is consistent with the 254

expected rate of 10-15% mutagenesis using equimolar concentrations of altered and natural 255

dNTPs (25). A higher degree of mutations can span a wider range in downstream 256

expression; however, it can also abolish the expression by altering important motifs for 257

transcriptional factor binding. During the in vitro growth the selected M. smegmatis::pJK 258

mutants showed a stable level of fluorescence. The PHsp60 promoter placed in the same 259

plasmid backbone as the pJK series induced a fluorescence intensity comparable to the 260

weakest PL5 promoter, thus indicating the high strength of PL5. A heat-shock treatment did 261

not increase the fluorescence of M. smegmatis::pJK or pJHsp60 (not shown). It could be 262

argued that fluorescence may not serve as a good parameter to determine protein expression 263

and ultimately promoter activity. However, here we have demonstrated that the 264

fluorescence determined by flow cytometry correlates with the total amount of GFP 265

expression in SDS-PAGE even though only 6 promoters were used. 266

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

14

In this study, we observed mutations throughout the whole sequence, including the -267

10 and -35 regions for the promoters, exhibiting high, medium or low strength. 268

Interestingly, all mid-strength promoters (pJK-B7, -E6 and -D6) showed a mutation at the 269

first position of the -10 region (C → T). Also, the weakest promoter that we identified, in 270

pJK-C1, is the only one with a mutation at the predicted transcriptional starting site (A → 271

G). 272

Heterologous expression in recombinant bacteria can be deleterious or lethal. Given 273

the mycobacterial relatively slow growth rate, where faster-growing strains like M. 274

smegmatis and M. fortuitum have a doubling-time of 2-3 hours and M. bovis BCG and M. 275

tuberculosis 20-24 hours (33), the eGFP synthesis undoubtedly represents a metabolic 276

burden and stress through consumption of ATP, nucleotides, aminoacids and usage of 277

ribosomes and tRNAs that may displace the expression of endogenous proteins and hence 278

inhibit the bacillus’s growth. GFP itself can also be toxic to the host (26). Nevertheless, in 279

this study, the only strain significantly different from the growth of the wild-type was M. 280

smegmatis::pJK-C1. This strain appears to have a lag in early time points but reaching a 281

similar O.D. at later time points. However, it is unlikely that a low and not a high 282

expression of eGFP could inhibit growth as observed by others (34). 283

To evaluate the utility of PL5 promoters for driving foreign antigen expression in 284

BCG, the sm29 gene from S. mansoni was cloned in place of the egfp gene downstream of a 285

strong (pJK-F8), an intermediate (pJK-B7) and a weak (pJK-C1) mutant promoter. 286

Recombinant BCG expressed Sm29 at different levels, which corresponded to their eGFP-287

measured promoter strength. pJK-F8 was not evaluated since no colonies of rBCG were 288

obtained even after several attempts. 289

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

15

rBCG::pJK-B7 clones were 9-fold more fluorescent than rBCG::pJK-C1 clones 290

(average FITC value of 27,300 and 3,200, respectively). To achieve such a difference, 291

Bourn used an enrichment method to select for high-copy number variants of pORI101 292

plasmid which were able to express GFP with a 7-fold increase compared to the original 293

(16). Although this strategy can be interesting to obtain a significant expression of cloned 294

genes, some disadvantages such as metabolic burden and plasmid instability may reduce 295

the usefulness of the system. 296

Since pJK plasmids are extrachromossomal copy number could be partially 297

responsible for the differences observed in fluorescence or Sm29 expression among 298

mycobacterial hosts. The origin of replication of pJH-egfp and pJK is derived from 299

pAL5000 of M. fortuitum, which is widely used in mycobacterial shuttle-plasmids such as 300

pMV261, pLA71 and pMIP12. Its copy number in M. smegmatis was determined as three 301

by Southern Blot (35), eight by antibiotic resistance (36) and twenty three by quantitative 302

PCR (37). The variation in the copy number between these studies can be explained by the 303

different methods of quantification, but also by the differences in plasmid structure. Despite 304

using the same pAL5000-derived origin of replication, other features such as resistance 305

marker, size and some other elements are also different. In our study all features are same, 306

with exception of a few nucleotides at the promoter sequence. Thus, we expect to minimize 307

the effect of plasmid copy number in the differential expression. Besides that, although 308

there are minor variations in the fluorescence during growth of recombinant M. smegmatis 309

their fluorescence values do not intersect with each other. 310

In conclusion, this study demonstrates the utility of the rational design of systems to 311

obtain a range of expression of target proteins in mycobacteria. The systems described 312

present an alternative approach to obtain expression of foreign genes in mycobacteria and 313

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

16

to design a new generation of recombinant BCG vaccines that optimize the host immune 314

response towards protection. 315

316

Acknowledgments 317

318

This work was supported by FAPESP grant 2008/04631-7, Santander through its program 319

Santander Universities and Fundação Butantan. 320

321

We thank S.T. Cole for the generous gift of pJH152 and G.F. Hatfull for pBRL8. 322

323

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

17

References 324

325

1. Bloom BR, Snapper SB, Kieser T, Jacobs WR. 1990. Development of 326 recombinant BCG vaccines. Seminars in Virology 1:7. 327

2. Bastos RG, Borsuk S, Seixas FK, Dellagostin OA. 2009. Recombinant 328 Mycobacterium bovis BCG. Vaccine 27:6495-6503. 329

3. Fuerst TR, de la Cruz VF, Bansal GP, Stover CK. 1992. Development and 330 analysis of recombinant BCG vector systems. AIDS Res Hum Retroviruses 331 8:1451-1455. 332

4. Ohara N, Yamada T. 2001. Recombinant BCG vaccines. Vaccine 19:4089-333 4098. 334

5. Dale JW, Dellagostin OA, Norman E, Barret ADT, McFadden J. 1993. 335 Multivalent BCG vaccines. In D. T. O'Hagan e (ed), Novel delivery systems for 336 oral vaccines. CRC Press, Boca Raton. 337

6. Dhar N, Rao V, Tyagi AK. 2004. Immunogenicity of recombinant BCG vaccine 338 strains overexpressing components of the antigen 85 complex of Mycobacterium 339 tuberculosis. Med Microbiol Immunol 193:19-25. 340

7. Dhar N, Rao V, Tyagi AK. 2003. Skewing of the Th1/Th2 responses in mice due 341 to variation in the level of expression of an antigen in a recombinant BCG system. 342 Immunol Lett 88:175-184. 343

8. Galen JE, Levine MM. 2001. Can a 'flawless' live vector vaccine strain be 344 engineered? Trends Microbiol 9:372-376. 345

9. Edelman R, Palmer K, Russ KG, Secrest HP, Becker JA, Bodison SA, Perry 346 JG, Sills AR, Barbour AG, Luke CJ, Hanson MS, Stover CK, Burlein JE, 347 Bansal GP, Connor EM, Koenig S. 1999. Safety and immunogenicity of 348 recombinant Bacille Calmette-Guerin (rBCG) expressing Borrelia burgdorferi 349 outer surface protein A (OspA) lipoprotein in adult volunteers: a candidate Lyme 350 disease vaccine. Vaccine 17:904-914. 351

10. Varaldo PB, Leite LC, Dias WO, Miyaji EN, Torres FI, Gebara VC, Armoa 352 GR, Campos AS, Matos DC, Winter N, Gicquel B, Vilar MM, McFadden J, 353 Almeida MS, Tendler M, McIntosh D. 2004. Recombinant Mycobacterium 354 bovis BCG expressing the Sm14 antigen of Schistosoma mansoni protects mice 355 from cercarial challenge. Infect Immun 72:3336-3343. 356

11. Ehrt S, Guo XV, Hickey CM, Ryou M, Monteleone M, Riley LW, 357 Schnappinger D. 2005. Controlling gene expression in mycobacteria with 358 anhydrotetracycline and Tet repressor. Nucleic Acids Res 33:e21. 359

12. Kanekiyo M, Matsuo K, Hamatake M, Hamano T, Ohsu T, Matsumoto S, 360 Yamada T, Yamazaki S, Hasegawa A, Yamamoto N, Honda M. 2005. 361 Mycobacterial codon optimization enhances antigen expression and virus-specific 362 immune responses in recombinant Mycobacterium bovis bacille Calmette-Guerin 363 expressing human immunodeficiency virus type 1 Gag. J Virol 79:8716-8723. 364

13. Varaldo PB, Miyaji EN, Vilar MM, Campos AS, Dias WO, Armoa GR, 365 Tendler M, Leite LC, McIntosh D. 2006. Mycobacterial codon optimization of 366 the gene encoding the Sm14 antigen of Schistosoma mansoni in recombinant 367

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

18

Mycobacterium bovis Bacille Calmette-Guerin enhances protein expression but 368 not protection against cercarial challenge in mice. FEMS Immunol Med 369 Microbiol 48:132-139. 370

14. Fan XY, Ma H, Guo J, Li ZM, Cheng ZH, Guo SQ, Zhao GP. 2009. A novel 371 differential expression system for gene modulation in Mycobacteria. Plasmid 372 61:39-46. 373

15. Andreu N, Zelmer A, Fletcher T, Elkington PT, Ward TH, Ripoll J, Parish 374 T, Bancroft GJ, Schaible U, Robertson BD, Wiles S. 2010. Optimisation of 375 bioluminescent reporters for use with mycobacteria. PLoS One 5:e10777. 376

16. Bourn WR, Jansen Y, Stutz H, Warren RM, Williamson AL, van Helden PD. 377 2007. Creation and characterisation of a high-copy-number version of the 378 pAL5000 mycobacterial replicon. Tuberculosis (Edinb) 87:481-488. 379

17. Hatfull GF, Sarkis GJ. 1993. DNA sequence, structure and gene expression of 380 mycobacteriophage L5: a phage system for mycobacterial genetics. Mol 381 Microbiol 7:395-405. 382

18. Nesbit CE, Levin ME, Donnelly-Wu MK, Hatfull GF. 1995. Transcriptional 383 regulation of repressor synthesis in mycobacteriophage L5. Mol Microbiol 384 17:1045-1056. 385

19. Taylor RG, Walker DC, Mcinnes RR. 1993. Escherichia-Coli Host Strains 386 Significantly Affect the Quality of Small-Scale Plasmid DNA Preparations Used 387 for Sequencing. Nucleic Acids Res 21:1677-1678. 388

20. Snapper SB, Melton RE, Mustafa S, Kieser T, Jacobs WR, Jr. 1990. Isolation 389 and characterization of efficient plasmid transformation mutants of 390 Mycobacterium smegmatis. Mol Microbiol 4:1911-1919. 391

21. Brosch R, Gordon SV, Garnier T, Eiglmeier K, Frigui W, Valenti P, Dos 392 Santos S, Duthoy S, Lacroix C, Garcia-Pelayo C, Inwald JK, Golby P, Garcia 393 JN, Hewinson RG, Behr MA, Quail MA, Churcher C, Barrell BG, Parkhill J, 394 Cole ST. 2007. Genome plasticity of BCG and impact on vaccine efficacy. Proc 395 Natl Acad Sci U S A 104:5596-5601. 396

22. Parish T, Stoker NG. 1998. Electroporation of mycobacteria. Methods Mol Biol 397 101:129-144. 398

23. Cardoso FC, Macedo GC, Gava E, Kitten GT, Mati VL, de Melo AL, Caliari 399 MV, Almeida GT, Venancio TM, Verjovski-Almeida S, Oliveira SC. 2008. 400 Schistosoma mansoni tegument protein Sm29 is able to induce a Th1-type of 401 immune response and protection against parasite infection. PLoS Negl Trop Dis 402 2:e308. 403

24. Cardoso FC, Pacifico RN, Mortara RA, Oliveira SC. 2006. Human antibody 404 responses of patients living in endemic areas for schistosomiasis to the tegumental 405 protein Sm29 identified through genomic studies. Clin Exp Immunol 144:382-406 391. 407

25. Zaccolo M, Williams DM, Brown DM, Gherardi E. 1996. An approach to 408 random mutagenesis of DNA using mixtures of triphosphate derivatives of 409 nucleoside analogues. J Mol Biol 255:589-603. 410

26. Alper H, Fischer C, Nevoigt E, Stephanopoulos G. 2005. Tuning genetic 411 control through promoter engineering. Proc Natl Acad Sci U S A 102:12678-412 12683. 413

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

19

27. Davis JH, Rubin AJ, Sauer RT. 2011. Design, construction and characterization 414 of a set of insulated bacterial promoters. Nucleic Acids Res 39:1131-1141. 415

28. Qin JY, Zhang L, Clift KL, Hulur I, Xiang AP, Ren BZ, Lahn BT. 2010. 416 Systematic comparison of constitutive promoters and the doxycycline-inducible 417 promoter. PLoS One 5:e10611. 418

29. Gall K, Barker LP. 2006. Differential green fluorescent protein expression from 419 mycobacterial promoter constructs in Escherichia coli and Mycobacterium 420 marinum. FEMS Microbiol Lett 255:301-307. 421

30. Kremer L, Baulard A, Estaquier J, Poulain-Godefroy O, Locht C. 1995. 422 Green fluorescent protein as a new expression marker in mycobacteria. Mol 423 Microbiol 17:913-922. 424

31. Muller-Taubenberger A, Anderson KI. 2007. Recent advances using green and 425 red fluorescent protein variants. Appl Microbiol Biotechnol 77:1-12. 426

32. Triccas JA, Britton WJ, Gicquel B. 2001. Isolation of strong expression signals 427 of Mycobacterium tuberculosis. Microbiology 147:1253-1258. 428

33. Dziadek J, Madiraju MV, Rutherford SA, Atkinson MA, Rajagopalan M. 429 2002. Physiological consequences associated with overproduction of 430 Mycobacterium tuberculosis FtsZ in mycobacterial hosts. Microbiology 148:961-431 971. 432

34. Luo Y, Szilvasi A, Chen X, DeWolf WC, O'Donnell MA. 1996. A novel 433 method for monitoring Mycobacterium bovis BCG trafficking with recombinant 434 BCG expressing green fluorescent protein. Clin Diagn Lab Immunol 3:761-768. 435

35. Ranes MG, Rauzier J, Lagranderie M, Gheorghiu M, Gicquel B. 1990. 436 Functional analysis of pAL5000, a plasmid from Mycobacterium fortuitum: 437 construction of a "mini" mycobacterium-Escherichia coli shuttle vector. J 438 Bacteriol 172:2793-2797. 439

36. Gavigan JA, Ainsa JA, Perez E, Otal I, Martin C. 1997. Isolation by genetic 440 labeling of a new mycobacterial plasmid, pJAZ38, from Mycobacterium 441 fortuitum. J Bacteriol 179:4115-4122. 442

37. Huff J, Czyz A, Landick R, Niederweis M. 2010. Taking phage integration to 443 the next level as a genetic tool for mycobacteria. Gene 468:8-19. 444

445

446

447

448

449

450

451

452

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

453

Figure 1. Schematic representation of pJH-egfp and the methodology for the 454

generation of a functional promoter library. (A) pJH-egfp contains the PL5 promoter 455

upstream the egfp sequence, a kanamycin resistance marker Kan(R), and origins of 456

replication in E. coli (oriC) and mycobacteria (oriM). (B) The PL5 promoter was 457

submitted to error-prone PCR and used to drive the expression of egfp. M. smegmatis 458

transformants were selected by kanamycin resistance and functional promoters screened 459

within fluorescent colonies. Selected colonies spanning a wide range of fluorescence 460

were further analyzed by microscopy, specific promoter mutations and single cell 461

fluorescence by flow cytometry (Adapted from (26)). 462

463

464

Figure 2. Screening of PL5 promoter mutants strength. (A) Clones of M. 465

smegmatis::pJH-egfp with the non-mutated PL5 promoter show a correlation between 466

growth (O.D.620) and fluorescence (F.U.log10) (r = 0.79). (B) M. smegmatis transformed 467

with pJK plasmid library show a variation in fluorescence independently of growth (r = 468

0.26). Wild-type M. smegmatis (no egfp gene) and M. smegmatis::pJH-egfp are indicated 469

by a black triangle and a star, respectively. The mutants selected for further analysis are 470

indicated as black circles. Statistical analysis used Two-tailed Spearman’s Rank 471

Correlation. 472

473

474

20

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

21

Figure 3. Sequence alignment of wild-type and mutated PL5 promoters in pJK 475

plasmids. The promoter sequence containing the selected pJK plasmids was sequenced 476

and compared to the original. From 257 bp comprising the promoter region, there are 6 – 477

11% of altered bp within mutants, which also localized to the predicted -10 and -35 478

regions. pJK-C1 showed a mutation at the transcriptional starting site (TSS). The 479

initiation codon is the last 3 bp in the sequence showed. 480

481

482

Figure 4. The error-prone PCR change the strength of PL5 promoter. (A) Relative 483

fluorescence (R.F.U.) of selected transformants comprising the fluorescence range. (B) 484

Microscopy of M. smegmatis::pJK demonstrate the differential fluorescence as referred 485

of promoter strength with little variation among the bacilli population. Statistical analysis 486

by one-way ANOVA (*** p < 0.001, ns = not significant) 487

488

489

Figure 5. pJK plasmid series induce differential expression in M. smegmatis. Late 490

log-phase cultures of M. smegmatis were transformed with selected pJK plasmids or 491

pJHsp60. (A) Fluorescence of recombinant M. smegmatis harboring pJK plasmid series 492

analyzed by flow cytometry. (B) Relative density of the GFP band in the SDS-PAGE 493

analyzed by densitometry. Statistical analysis by one-way ANOVA (*** p < 0.001, ** p 494

< 0.01, ns = not significant). 495

496

497

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

22

Figure 6. pJK plasmids series induce differential fluorescence in M. bovis BCG. Late 498

log-phase cultures of BCG were transformed with pJK plasmid series or pJHsp60 and 499

fluorescence analyzed by flow cytometry. Statistical analysis by one-way ANOVA (*** 500

p < 0.001, ns = not significant). 501

502

503

Figure 7. Western blot of rBCG strains differentially expressing Sm29 antigen 504

through mutated PL5 promoters. Total protein extracts (25 µg) of wild-type BCG (Lane 505

2) or rBCG::pJK-C1.Sm29 (Lanes 3-4) and pJK-B7.Sm29 (Lanes 5-6). Two 506

independently-isolated clones were analyzed by Western blot using anti-rSm29 antibodies. 507

300 ng of rSm29 protein (Lane 1) was used as positive control. Molecular weight markers 508

standard are indicated on the left 509

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from

on April 7, 2018 by guest

http://aem.asm

.org/D

ownloaded from