Defining a temporal order of genetic ... - Wiley Online Library

Molecular Microbiology (2017) 105(5), 794–809 䊏

doi:10.1111/mmi.13734 First published online 7 July 2017

Defining a temporal order of genetic requirements for development of mycobacterial biofilms Yong Yang,1 Joseph Thomas,2 Yunlong Li,1 Catherine Vilche`ze,3 Keith M. Derbyshire,1 William R. Jacobs Jr3,4 and Anil K. Ojha 1,2* 1 Division of Genetics, Wadsworth Center, New York State Department of Health, Albany, NY, USA. 2 Department of Infectious Diseases and Microbiology, University of Pittsburgh, Pittsburgh, PA, USA. 3 Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, NY, USA. 4 Howard Hughes Medical Institute, Bronx, NY, USA.

Summary Most mycobacterial species spontaneously form biofilms, inducing unique growth physiologies and reducing drug sensitivity. Biofilm growth progresses through three genetically programmed stages: substratum attachment, intercellular aggregation and architecture maturation. Growth of Mycobacterium smegmatis biofilms requires multiple factors including a chaperonin (GroEL1) and a nucleoid-associated protein (Lsr2), although how their activities are linked remains unclear. Here it is shown that Lsr2 participates in intercellular aggregation, but substratum attachment of Lsr2 mutants is unaffected, thereby genetically distinguishing these developmental stages. Further, a suppressor mutation in a glycopeptidolipid synthesis gene (mps) that results in hyperaggregation of cells and fully restores the form and functions of Dlsr2 mutant biofilms was identified. Suppression by the mps mutation is specific to Dlsr2; it does not rescue the maturation-deficient biofilms of a DgroEL1 mutant, thereby differentiating the process of aggregation from maturation. Gene expression analysis supports a stepwise process of maturation, highlighted by temporally separated, transient inductions of iron and nitrogen import genes. Furthermore, GroEL1 activity is required for induction of nitrogen, but not iron, import genes. Together, the findings begin to define molecular Accepted 17 June, 2017. *For correspondence. E-mail anil.ojha@ health.ny.gov; Tel. 1412 648 8058; Fax 1412 383 8926. C 2017 John Wiley & Sons Ltd V

checkpoints during development of mycobacterial biofilms.

Introduction Most microbial species naturally grow as sessile, matrixencapsulated, multicellular communities, called biofilms, which attach to surfaces or air–liquid interfaces (Fux et al., 2005; Lopez et al., 2010; Islam et al., 2012). Biofilm growth is a programmed process, in which participating cells must establish physical contacts and chemical communications among each other, and constantly adapt to the changing microenvironments within the growing three-dimensional architecture (Davies et al., 1998; Hogan and Kolter, 2002; Camilli and Bassler, 2006; Sakuragi and Kolter, 2007). The varied growth dynamics within biofilms result in considerable cellular heterogeneity, which along with the physical protection from the extracellular matrix, promotes bacterial persistence under exogenous stress (Stoodley et al., 2002; Davies, 2003). In the human body, a vast majority of microbial infections are considered to occur as biofilms, which are extraordinarily recalcitrant to antimicrobial therapeutics (Mah and O’Toole, 2001; Davies, 2003; Fux et al., 2005; Hoiby et al., 2010). Biofilms also constitute a unique host–pathogen interface, in which the resident microbes manipulate and evade host immunity to facilitate their persistence (Leid et al., 2005; Anderson and O’Toole, 2008; Prabhakara et al., 2011). The assemblage of microbes into biofilms proceeds through distinct stages. These include the attachment of single-cell planktonic microbes onto a substratum; aggregation and growth of the adherent cells into threedimensionally organized structures; and encapsulation of the structures by a self-produced matrix of extracellular polymeric substance. Each of these stages, defined by specific genetic requirements, is hierarchically ordered in the developmental processes (Pratt and Kolter, 1998; Stanley and Lazazzera, 2004; Moorthy and Watnick, 2005; Petrova and Sauer, 2009). For example, Vibrio cholerae biofilm development requires cheY-3 dependent monolayer formation, which is subsequently followed by bap1 and leuO dependent matrix synthesis

Developmental stages of mycobacterial biofilms 795

(Moorthy and Watnick, 2005). Similarly, biofilm development in Pseudomonas aeruginosa occurs in four stages, initiated by reversible attachment that proceeds to irreversible attachment, followed by two stages of maturation (Petrova and Sauer, 2009). While bfiS is required for transition from reversible to irreversible stage, bfmR facilitates transition from irreversible attachment to maturation-1 stage, and mifR appears to be involved in transition from maturation-1 to maturation-2 stage (Petrova and Sauer, 2009). Anchorage of Staphylococci biofilms to the host substrata is facilitated by MSCRAMMs (microbial surface components recognizing adhesive matrix molecules), whereas intercellular aggregation requires the presence of the ica locus encoding polysaccharide intercellular adhesin (PIA) and accumulation associated proteins Aap and SasG (Heilmann et al., 1996; Cramton et al., 1999; Cucarella et al., 2001; Conrady et al., 2008). Pathogenic and environmental species of mycobacteria including Mycobacterium tuberculosis, Mycobacterium avium, Mycobacterium ulcerans and Mycobacterium smegmatis form drug-tolerant biofilms through dedicated genetic programs (Ojha et al., 2005; Hall-Stoodley et al., 2006; Yamazaki et al., 2006; Marsollier et al., 2007; Ojha et al., 2008; Bosio et al., 2012). Moreover, a multi-stage developmental process of mycobacterial biofilms is suggested by the arrest of a DgroEL1 mutant of M. smegmatis at the monolayer stage (Ojha et al., 2005, 2010). However, molecular demarcations of the developmental stages and the underlying checkpoints remain unclear for mycobacterial biofilms. In this study, fortuitous discovery of a suppressor of a biofilm-defective mutant, Dlsr2, has allowed us to address this gap. By analysis of the suppressor, we distinguish substratum attachment from intercellular aggregation in mycobacterial biofilms, and show that intercellular aggregation but not substratum attachment is necessary for subsequent development of matured biofilms. The findings reveal a hierarchical order of developmental stages in mycobacterial biofilm formation.

Results Suppression of a Dlsr2 mutant by a mutation in mps To isolate biofilm-defective mutants, a pool of 100,000 transposon mutants of M. smegmatis mc2155 was cultured in a syringe containing detergent-free medium (Supporting Information Fig. S1). Thick pellicles developed at the air–liquid interface after 5 days of incubation. We hypothesized that the liquid underlying the pellicle contained mutants, which either existed predominantly in planktonic form, or were micro-clusters of cells C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

dislodged from the pellicles. To enrich and isolate this subpopulation, we filtered the contents of the syringe through a 5 lm filter. We recovered 120 colonies representing 27 unique mutants from the filtrate (Supporting Information Table S1). Of these, 28 had an insertion at the 449th nucleotide of MSMEG_2952, and 61 had an insertion at the 261st nucleotide of MSMEG_6092, which encodes for a nucleoid associated protein, Lsr2. While the isolation of lsr2 as a biofilm-defective mutant is consistent with previous reports (Chen et al., 2006; Nguyen et al., 2010), the abundance of siblings of lsr2 and MSMEG_2952 mutants in the filtrate suggests that the mutants are able to grow better in the relatively hypoxic condition below the pellicles. To further understand their phenotypes, we constructed precise deletions of MSMEG_2952 and lsr2 in mc2155. Both mutants failed to grow at the air–medium interface, instead accumulated biomass at the bottom of the culture (Fig. 1A,B). This was corroborated by increase in CFU of DMSMEG_2952 and Dlsr2 among the submerged subpopulation (Supporting Information Fig. S2A). Moreover, both mutants formed smooth colonies on a solid medium, but exhibited normal growth in planktonic culture (Supporting Information Fig. S2B,C). During unsuccessful attempts to complement DMSMEG_2952, we obtained three rough-colony revertants that maintained their morphology even in the absence of the complementing plasmid (Supporting Information Fig. S3A). Because of the deletion in MSMEG_2952 (Supporting Information Fig. S3B), the smooth-to-rough colony conversion was interpreted to be due to an extragenic suppressor mutation. Wholegenome sequence analysis of one of the revertants, followed by targeted PCR of all three clones, revealed a single nucleotide insertion in MSMEG_0400 gene that truncates its open reading frame (data not shown). MSMEG_0400 encodes a mycobacterial non-ribosomal peptide synthase (mps), which synthesizes the peptide backbone of glycopeptidolipids (GPLs) (Billman-Jacobe et al., 1999). Mutations in mps have been shown to eliminate the outermost capsular layer of the envelope, reduce cell permeability, cause hyperaggregation, and result in a colony morphology rougher than mc2155 (Billman-Jacobe et al., 1999; Etienne et al., 2002). To verify the suppressor mutation, we engineered an amber mutation at the 11th residue (S11X) of mps in an unmarked DMSMEG_2952 strain (Supporting Information Fig. S3C). The double mutant, called DMSMEG_2952/ mpsS11X, was expectedly devoid of GPL and formed fully matured biofilms that were indistinguishable from the parent mc2155 (Fig. 1A–C). Similar to DMSMEG_2952/ mpsS11X, a deletion of mps in a Dlsr2 background (Dlsr2/ Dmps) resulted in loss of GPLs and substantially restored biofilm formation (Fig. 1A, B and D).

䊏

796 Y. Yang et al.

A

mc2155

MSMEG_2952 EG G 295 52

lsr2

lsr2comp

lsr2/

mps

MSMEG_2952/ mpsS11X

groEL1

groEL1/

mps

mps

C

D

B *** ***

0.06 0.04 0.02

m ps

gr g oE roE L1 L1 / m ps

1X

ND

ls M m SM p M r2/ EG SM m _2 EG ps 95 _ 2/ 29 m 52 ps S 1

co

ls r2

ND

r2

m

c2 15 5

0.00

ls

pellicle biobass (mg / mL)

0.08

Interestingly, deletion of mps did not rescue the impaired biofilms of a DgroEL1 mutant (Ojha et al., 2005), implying distinct roles for Lsr2 and GroEL1 in formation of biofilms (Fig. 1A,B).

Cell-substratum attachment and intercellular aggregation have distinct genetic requirements Suppression of Dlsr2 and DMSMEG_2952 by an mps mutation was surprising, given that GPLs have been C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


Fig. 1. Suppression of biofilm-defective mutants, Dlsr2 and DMSMEG_2952, by a secondary mutation in a GPL biosynthesis gene, mps. A. Top–down view of pellicle biofilms of the indicated strains formed on an air–liquid interface after 5 days of incubation. mc2155 is the parent strain of M. smegmatis. Dlsr2/Dmps and Dlsr2/DgroEL1 denote deletions of both genes in the same strain; DMSMEG_2952/mpsS11X denotes truncation of mps in DMSMEG_2952 strain; Dlsr2comp denotes a Dlsr2 strain complemented with a plasmid carrying the wild-type copy of lsr2. While growth in Dlsr2 and DMSMEG_2952 occurs beneath the liquid at the bottom on the culture, DgroEL1 and DgroEL1/Dmps form untextured monolayers at the interface that contrast with the textured pellicle produced by the other strains. B. Biomass (wet weight per ml of culture medium) of 5-day-old pellicles from the indicated strains (mean 6 SD, n 5 3, ***p < 0.001). Values for Dlsr2 and DMSMEG_2952 could not be determined (ND) due to lack of pellicles. C and D. Thin layer chromatography (TLC) showing loss of glycopeptidolipids (GPL) in DMSMEG_2952/mpsS11X (C), and Dlsr2/Dmps (D). mc2155 and the respective single mutant strains were used as controls in both panels. The arrows indicate four major species of apolar GPL constitutively present in the envelope, which differ from each other with respect to the extent of methylation on the sugar moieties (Ojha et al., 2002).

proposed to have an important role in M. smegmatis biofilm formation by facilitating cellular attachment to the substratum and sliding motility (Recht and Kolter, 2001). To further clarify this apparent discrepancy, we analyzed biofilms of a deletion mutant of mps in mc2155 background. The mutant was confirmed for loss of GPL in its envelope (Supporting Information Fig. S3D). Similar to mc2155, the Dmps strain formed biofilms at the air– medium interface (Fig. 1A), which together with earlier reports of the hyperaggregative and rough-colony morphology of a Dmps strain (Billman-Jacobe et al., 1999; Etienne et al., 2002), led us to hypothesize that the mutants from our screen would be defective in intercellular aggregation. We therefore evaluated the aggregation proficiencies of DMSMEG_2952, Dlsr2 and their suppressor strains. We determined the rate of decrease in optical density of cells in a standing, detergent-free liquid suspension. In this assay, reduction in cumulative cell surface area upon aggregation proportionately reduces the total light scatter (Malik et al., 2003). Aggregation of a suspension of wild-type and Dlsr2 complemented cells in detergent-free PBS progressively increased to a maximum value in about 10 min, while aggregation of Dlsr2 and DMSMEG_2952 cells was significantly delayed (Fig. 2A). However, aggregation of the suppressor strains reached a saturating value in < 5 min, consistent with the previously observed hyperaggregation with Dmps cells (Fig. 2A). The findings of the optical assay were further verified by a qualitative microscopic assay, in which 200 ll of homogenously suspended cells of a fluorescently labeled strain was placed on a pre-focused glass-bottom dish on an inverted microscope and appearance of fluorescent aggregates at the bottom was measured over a 10-min period. Consistent with the optical assay, Dlsr2 produced fewer aggregates than mc2155, but numerous large aggregates of Dlsr2/Dmps were easily evident in the period (Fig. 2B). We further investigated if substratum attachment of cells was impacted by the mutations. Detergent-free suspensions of fluorescently labeled strains were analyzed for their attachment to a polystyrene substratum C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

after 24 h of incubation at 378C. In this assay, both DMSMEG_2952 and Dlsr2 cells formed a uniform monolayer on the substratum, similar to mc2155 (Fig. 2C,D). By contrast, attachment of their suppressor strains was non-uniform, and dominated by cellular aggregates (Fig. 2C,D). This implies that substratum attachment and intercellular aggregation require distinct surface properties of the envelope, determined by different sets of genes. Furthermore, GPL appears to have opposite role in substratum attachment and intercellular aggregation. Importantly, neither substratum attachment nor intercellular aggregation was impaired in the DgroEL1 mutant (Supporting Information Fig. S4): consistent with a role of the chaperonin at a developmental stage following attachment and intercellular aggregation. To precisely define the role of GPL in biofilm development, we closely compared the growth and development of pellicle biofilms by Dmps with mc2155. Expectedly, mc2155 formed a uniform monolayer along the sidewall of the container as well as on the air–medium interface by 3rd day of incubation, developing into matured pellicles by 5th day (Fig. 3A,B). By contrast, Dmps mutant primarily formed suspended and isolated aggregates, with spotted attachment foci on the sidewall after 3 days of incubation (Fig. 3A,B). Although the biomass of the suspended aggregates of Dmps increased by the 4th day, unoccupied space on the air–medium interface remained evident (Fig. 3B). However, the mutant formed a fully textured pellicle, similar to its parent mc2155, by the 5th day. The developmental course of Dmps biofilms is consistent with a specific role of GPL as a facilitator of cellular motility and subsequent spreading on a surface, likely by reducing cell–cell aggregation. Phenotypic similarity between DMSMEG_2952 and Dlsr2 led us to analyze the expression of each of the two genes upon mutation in the other. While expression of MSMEG_2952 remained unchanged upon lsr2 mutation, expression of lsr2 was about eightfold less in DMSMEG_2952 (Supporting Information Fig. S5). Moreover, overexpression of lsr2 from a constitutively strong promoter (Phsp60) restored the biofilm development of DMSMEG_2952 (Supporting Information Fig. S5). It is

798 Y. Yang et al.

䊏

A

B mc2155

aggregation index

100

0 min

10min

10µm

10µm

10µm

10µm

10µm

10µm

2

mc 155 lsr2comp lsr2 lsr2/ mps mps lsr2 MSMEG_2952 MSMEG_2952/mpsS11X

80 60 40 20

lsr2/ mps

0 0

5 time (minute)

10

C MSMEG_2952

mc2155

D

% area coverage

50

ns ns

lsr2

MSMEG_2952/mpsS11X

lsr2/ mps

**** ****

40 30 20 10

M mc 2 SM EG 155 M SM _2 EG 95 2 _2 95 ls 2/ r2 m ps S 11 ls X r2 / m ps

0

Fig. 2. Genetic distinction between substratum attachment and intercellular aggregation. A. Aggregation index of the indicated strains suspended in detergent-free PBS over the indicated time period. The strain nomenclature is same as described for Fig. 1A. B. Visualization of aggregates of the indicated strains (expressing mCherry) settled at the bottom of a glass surface from a suspension over a 10-min period. C. A representative fluorescent micrograph showing the indicated strains, each expressing mCherry constitutively, attached to the bottom of a polystyrene plate after 24-h of incubation. The plate was washed three times to remove non-adherent cells before imaging. D. The plot shows the coverage of the surface by each strain in five biologically independent micrographs (mean 6 SD, n 5 5, **** denotes p < 0.0001 by unpaired t-test). The values, obtained using ImageJ software, represent the ratios of the total pixels associated with bright areas to the total area. C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


A

mc2155

mps

(3-day)

B

mc2155

mps

(Fig. 4A,B). We further tested Dlsr2/Dmps biofilms for tolerance to antibiotics and conjugal transfer of DNA, two established properties of mycobacterial biofilms (Nguyen et al., 2010; Ojha et al., 2010). A substantial restoration of both phenotypes was observed in Dlsr2/ Dmps biofilms (Fig. 4C,D). Together, these three assays indicate that the biofilms formed by the suppressor are morphologically and functionally similar to that of the parent mc2155 strain.

Lsr2 has an indirect role in biofilm formation (3-day)

(4-day)

(5-day)

Fig. 3. Biofilm development by mc2155 and Dmps strains. A. A side view of biofilms of the two strains formed on the sidewall of the culture dish after 3-day incubation. The arrow indicates the horizontal plane of the air–medium interface. B. A top–down view of pellicle biofilms of the two strains across the air–medium interface at the indicated time point.

thus likely that lsr2 expression is responsive to a cis-element in MSMEG_2952. To avoid redundancy, we omitted DMSMEG_2952 and its suppressor from further studies.

Biofilms of Dlsr2/Dmps have similar properties to mc 2155 biofilms We next investigated whether the Dlsr2/Dmps strain exhibited other characteristics associated with the biofilms of mc2155. The extracellular abundance of free mycolic acids (FM), as a matrix material, is one of the key structural features of mycobacterial biofilms (Ojha et al., 2008, 2010). While the FM levels were reduced in Dlsr2 biofilms, the levels in Dlsr2/Dmps biofilms were similar to mc2155, suggesting a restoration of signals that trigger FM synthesis and maturation of biofilms C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

Lsr2 binds at hundreds of genomic sites and impacts the expression of a large set of genes in M. smegmatis (Colangeli et al., 2007; Gordon et al., 2010). Thus, lsr2 mutations have pleiotropic effects that complicate the interpretation of the mutant phenotypes. The suppressor of Dlsr2 offers an advance. We reasoned that genes that have lower expression in Dlsr2, but restored back to the mc2155 level in Dlsr2/Dmps, could have consequential role in morphological and/or functional development of biofilms. We therefore compared the transcriptomes of mc2155, Dlsr2 and Dlsr2/Dmps strains grown in biofilms using RNA-seq (Supporting Information Table S2). Because the Dlsr2 mutant does not form pellicle biofilms, but can be distinguished from mc2155 and Dlsr2/ Dmps as colonies, we assessed colony biofilm as a model to compare the three strains. The transcriptomes from mc2155 and Dlsr2 colonies identified 162 genes with increased expression (Fig. 5A, Supporting Information Table S3), and 112 genes with decreased expression in the mutant (Fig. 5A, Supporting Information Table S4), when a statistically significant (q < 0.05) fourfold difference was used as the cut-off point. The differences observed from RNA-seq data are much greater than those identified in a previous microarray-based transcriptomic study performed on planktonic cultures of Dlsr2 and mc2155 (Colangeli et al., 2007). This could be attributed either to differences in growth condition or a greater linear dynamic range of the RNA-seq technique. However, similar to our data, more genes with increased expression than those with decreased expression were found in the previous study (Colangeli et al., 2007). This finding is in agreement with the idea that Lsr2 is primarily a negative regulator of gene expression. No difference in the expression of mps was observed (Supporting Information Table S2), which was consistent with the similar levels of GPLs found in both Dlsr2 and mc2155 (Fig. 1C). However, Kocincova et al. proposed that Lsr2 negatively regulates mps expression (Kocincova et al., 2008). Their conclusion was based on analysis of smooth colony variants, which produced high levels of GPL and had

800 Y. Yang et al.

䊏

A

B

FM

% FM in apolar lipids

40

**

**

30 20 10

ls r2

ls r2 /

m c2 15 5 10-2

D *

mc2155 lsr2 lsr2/ mps

100

10-3

*

% survival

transfer frequency/ donor

C

10-4 10-5

m ps

0

10 1

m ps

ls r2 /

ls r2

m c2 15 5

10-6

0.1 0

2

4

6

8

exposure time (day)

Fig. 4. Physiological and functional similarities between mc2155 and the suppressor (Dlsr2/Dmps) biofilms. A. Radio-TLC showing free mycolic acid (FM) synthesized by mc2155, Dlsr2 and its suppressor Dlsr2/Dmps after 5 days of growth in biofilms. Purified 14C-FM was used as a marker (M). Cells were labeled with 14C-acetate, and radiolabeled lipids equivalent to 10 000 cpm were loaded for analysis by thin-layer chromatography in chloroform:methanol (96:4). B. Quantitative analysis of the level of FM as a percentage of total apolar lipids in Fig. 4A. Data represent mean 6 SE (n 5 3, denotes p < 0.01 by unpaired t-test). C. Frequency of conjugal transfer of DNA between a recipient strain of M. smegmatis MKD8, and the indicated donor strains (mean 6 SD, n 5 3, denotes p < 0.05 by unpaired t-test). D. Sensitivity of 5-day-old biofilms of the indicated strains to 400 lg ml21 of rifampicin. Following 5 days of growth, rifampicin was added in the liquid beneath the pellicles and incubated for the times indicated on the x-axis. At each time point, cells were harvested and plated on fresh medium to determine their viability (mean 6 SD, n 5 3).

mutations in lsr2. It is plausible that the consequence of Lsr2 on mps expression is conditional upon the presence of additional unidentified mutations in the variants. The transcriptome profile of the Dlsr2/Dmps suppressor cells showed a striking contrast between genes positively and negatively impacted by the loss of Lsr2 (Fig. 5A,B, Supporting Information Tables S3 and S4). Expression of all but one (except MSMEG_6451) genes negatively impacted by a deletion of lsr2 was restored by the secondary deletion of mps, implying that these genes are under indirect control of Lsr2 (Fig. 5A,B and

Supporting Information Table S4). A randomly selected subset of five genes confirmed the RNA-seq patterns (Fig. 5C). Deletion of MSMEG_6451 in mc2155 did not produce any detectable biofilm phenotype, thereby ruling out a significant contribution to the Dlsr2 phenotype (data not shown). Only seven genes with elevated expression in the Dlsr2 reverted back to mc2155-like expression in the Dlsr2/Dmps double mutant (Supporting Information Table S3). Deletion of these seven genes in a Dlsr2 background failed to alter the Lsr2-dependent deficiency in biofilm formation (data not shown), C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


A

B

1000

1000 100

-log10(qValue)

-log10(qValue)

100 10 1 0.1 0.01

10 1 0.1 0.01 0.001

0.001 0.0001

0.0001 -5

0

5

10

log2( lsr2 : mc2155)

-5

0

5

10

log2( lsr2/ mps : mc2155)

mc2155 lsr2 lsr2/ mps

C relative expression

3 1 -1 -3

M SM EG _0 05 1 M SM EG _0 91 8 M SM EG _1 08 2 M SM EG _3 99 7 M SM EG _5 21 4

-5

Fig. 5. Transcriptomics of colony cultures of mc2155, Dlsr2 and Dlsr2/Dmps strains on agar plates. A and B. Volcano plots depicting statistically significant differences in the transcript levels of genes between mc2155 and Dlsr2 (A), and mc2155 and Dlsr2/Dmps (B) strains obtained by RNA-seq. The y-axis shows the statistical significance (q-values) of the mRNA fold difference, shown on the x-axis. The values in green (increased) and red (decreased) boxes represent genes with a greater than fourfold, statistically significant (q < 0.05) differences in the expression. Complete list of these genes is in Supporting Information Table S2. C. Validation of the differential expression of a select set of genes (indicated on x-axis) by RT-qPCR (mean 6 SD, n 5 3). The sigA transcript was used as an endogenous reference for normalization.

suggesting that expression of these genes does not contribute to the Dlsr2 phenotype. Together, the transcriptome of the suppressor confirms Lsr2 as a negative regulator. Although the precise effect of Lsr2 on the envelope that determines aggregation remains unknown, it appears to be less important in a GPL-free envelope.

Nitrogen starvation response in late biofilm development To understand the significance of the 111 genes (excluding lsr2) whose expression is restored to the mc2155 levels upon deletion of mps in a Dlsr2 mutant, we examined using RNA-seq their expression in early (84 h post inoculation) and late (144 h post inoculation) C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

stages of mc2155 pellicle biofilms, compared with their expression in a planktonic culture (Supporting Information Table S5). Eighty-three of these genes were induced exclusively at the late stage, and one was induced at both stages, while the expression of the remaining 27 genes did not change significantly between the two conditions (Fig. 6A, Supporting Information Table S6). The late expression of the 83 genes is consistent with their induction after successful establishment of Lsr2-dependent aggregated growth. A majority of the 83 genes had been detected in a previous microarray-based transcriptomic study of M. smegmatis biofilms (Ojha and Hatfull, 2007). Their restored expression in a genetic suppressor of an aggregation-deficient mutant suggests they are induced in response to

䊏

802 Y. Yang et al.

A

B

Complete N0 C0

4

relative expression

2 0 -2 -4 -6 -8

C

mc2155

M SM EG _6 25 9

M SM EG _4 63 5

M SM EG _2 42 5

-10

2425-27

4635-36

6259-68 complete

0

5

10

N1/2

C1/2

D

MSMEG_6259-68 mc2155 MSMEG_2425-27 MSMEG_4635-36

2.0

E

4

OD600

relative expression

1.5 1.0 0.5 0.0 0.0

0.5

1.0

1.5

0 -2 -4 -6 -8 -10

2.0

3

time (day)

F

4

5

6

7

time (day) mc2155_plnk mc2155_bf(5d) groEL1_bf(5d)

2 0 -2 -4

M SM EG _6 25 9

M SM EG _4 63 5

-6

M SM EG _2 42 5

relative expression

MSMEG_2425 MSMEG_0615 MSMEG_4509

2

C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


Fig. 6. Genes with restored expression in the Dlsr2/Dmps suppressor are induced during maturation of mc2155 biofilms. A. Heat map depicting difference in expression levels of 111 genes, which are decreased in a Dlsr2 strain but restored in a Dlsr2/Dmps strain (Supporting Information Table S4). Expression of genes was measured in early and late biofilms (bf) of mc2155, and compared with their basal levels in a planktonic (plnk) culture of the strain. Eighty-three of these genes are expressed exclusively in late-stage biofilms (see the list in Supporting Information Table S6). B. Expression of three amt genes (amt1: MSMEG_ 2425, amt2: MSMEG_4635 and amt3: MSMEG_6259) in planktonic mc2155 cells after 3 h of incubation in N0, or C0 variations of biofilm (bf) medium, free of primary sources of nitrogen and carbon respectively. C. Five-day biofilms of the wild-type and indicated mutants in N1/2 or C1/2 variations of complete bf medium containing half the amounts of these nutrients. The mutants are denoted with their respective MSMEG_ numbers. D. planktonic growth of wild-type and mutants shown in panel C in N1/2 biofilm medium. E. Expression of Fe sequestering genes (MSMEG_0615 and MSMEG_4509) and amt1 during biofilm growth of mc2155 at the indicated time post inoculation. F. Comparison of expression of the amt genes in planktonic (plnk) and 5-day (5d) biofilms (bf) of wild-type and DgroEL1 cells. All expression levels were determined by RT-qPCR and Ct values were normalized using the sigA transcript. Data in panels B, D, E and F represent mean 6 SD, n 5 3.

intercellular aggregation, and thus unambiguously defines these 83 genes as biofilm-specific factors associated with the late-stage of development. About 51 out of the 83 genes, including putative nitrite reductases (MSMEG_0427-28), ammonium transporters (amt1: MSMEG_2425, amt2: MSMEG_4635 and amt3: MSMEG_6259) and Glu/Gln synthetases (MSMEG_6260 and 6263), are controlled by a nitrogen-starvation sensing regulator GlnR (Amon et al., 2008; Jenkins et al., 2013) (Supporting Information Table S6). This suggests a biofilm-specific N-starvation response in cells during late stages of development. To further confirm the Nstarvation response in late biofilms, we examined the effect of N depletion on expression of the amt genes in planktonic cultures of mc2155. The expression of all three amt genes was induced by 60- to 200-fold within 3 h of incubation of cells in N0 medium, which lacked supplemental ammonium sulfate and casamino acids, but had normal glucose (Fig. 6B). However, incubation in C0 medium (without any primary carbon source, but with normal levels of N sources) did not produce a transcriptional response in the genes (Fig. 6B). As expected, deletion mutants of loci containing amt genes, DMSMEG_2425–27, DMSMEG_4635–36 and DMSMEG_6259–68, failed to form mature biofilms in N1/2-medium (50% of normal levels of ammonium sulfate and casamino acids) (Fig. 6C). The biomass of the mutants were correspondingly lower than the wild type (Supporting Information Fig. S6A), although no significant difference in cell density was observed (Supporting Information Fig. S6B,C). This suggests that nitrogen assimilation in the mutants was above the threshold of the amount required for viability, but below the threshold necessary to produce biomass for the pellicles. This is further supported by the fact that the mutants showed no significant deficiency during planktonic growth in N1/2 medium (Fig. 6D). Consistent with the N-specific induction of the genes, the mutants formed normal biofilms in C1/2-medium (50% of normal levels of glucose) (Fig. 6C). C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

The amt genes are induced after Fe import genes in biofilms A biofilm-specific response to iron (Fe) limitation, reflected by induced expression of Fe sequestration pathways and their requirement for biofilm formation, has been described previously (Ojha and Hatfull, 2007). Interestingly, the RNA-seq data in this study suggested that the genes involved in Fe sequestration, including exochelin and mycobactin biosynthesis (MSMEG_001115, MSMEG_4509-16) (Ojha and Hatfull, 2007), as well as the ESX-3 system (MSMEG_0615-25) (Siegrist et al., 2009), are induced earlier than the 83 genes associated with late biofilms including the three amt genes (Supporting Information Table S6). This highlights distinct transcriptional programs in biofilm formation, defined by their specific gene expression patterns. To precisely determine the induction times of Fe and N import genes during biofilm development, we analyzed the expression levels of an amt gene (amt1) and two genes representing Fe sequestration systems (MSMEG_0615 and MSMEG_4509). The mRNA samples were collected from an early (day-3) stage through to the matured biofilm stage (day-7) (Ojha et al., 2005). Induction of the two Fe sequestration genes peaked at day-4, while induction of amt1 was maximal at day-5 (Fig. 6E). Expression of all three genes decreased to a basal level by day-6 (Fig. 6E), suggesting that the inductions of Fe and N import genes are driven more by increased metabolic demand during developmental phases of biofilm formation, rather than by acute starvation in the restrictive microenvironment within biofilms.

Induction of amt, but not Fe import genes, requires GroEL1 activity in biofilms The chaperonin GroEL1 modulates mycolic acid biosynthesis to produce FM during maturation of biofilms (Ojha et al., 2005, 2010). We therefore examined the requirement for GroEL1-dependent functions for induction of Fe and N import activities. Induction of the Fe

804 Y. Yang et al.

䊏 FM synthesis (GroEL1) N assimilation (Amt1 Amt2, Amt3, Glu/Gln synthetases)

Mps

q Fe sequestration (ESX-3, exochelin, mycobactin) planktonic

Lsr2

Mps

substratum b t t attachment

aggregated growth 0-2 days

sliding/ spreading

early maturation

late maturation

2-4 days

> 4 days

Fig. 7. A model depicting the cellular and molecular events and their approximate times that demarcate the stages of biofilm development in M. smegmatis. Levels of FM synthesis, Fe sequestration and N import/assimilation in early and late maturation stages, relative to early stage of aggregated growth, are shown above the developmental scheme. Representations are not scaled. The genes used in this study to define these stages are indicated. Planktonic cells attach to a substratum, the coverage of which is determined by Mps-dependent GPLs in the envelope. GPLs inhibit aggregation and also allow spreading of cells on the substratum. Lsr2 is required for aggregated growth of cells containing GPLs in the envelope. Growth of aggregates triggers GroEL1-dependent FM synthesis, and induction of Fe-sequestration pathways that mark the early stage of maturation. These are followed by induction of N-assimilation/import genes at late-stage of maturation.

import genes in the day-4 biofilms was mostly unaffected in a DgroEL1 mutant (Supporting Information Fig. S7A). However, induction of all three amt genes in day5 biofilms required GroEL1 activity (Fig. 6F). Their induction in a nitrogen-limiting planktonic culture of a DgroEL1 strain was unaffected (Supporting Information Fig. S7B), thereby ruling out a direct role of GroEL1 in their regulation. Thus, signals inducing amt genes, and perhaps other N import genes, are likely generated from a GroEL1-dependent checkpoint that likely occurs before the day-5 stage.

Discussion Our study sheds light on the likely order of molecular events and their genetic requirements during development of M. smegmatis biofilms (Fig. 7). During early stages of biofilm formation, Lsr2 appears to be critical for aggregated growth of cells that contain GPLs on the surface. However, the precise determination of the Lsr2dependent aggregation factor(s) requires further investigation. In contrast to the effect of Lsr2, GPLs in the envelope appear to inhibit aggregation or promote disaggregation. Thus, the presence of GPL in the envelope could maximize the surface area of the suspended planktonic aggregates during initial attachment to a substratum, and also allow spreading of the aggregates on the substratum by facilitating sliding motility in the loosely-held cells (Recht and Kolter, 2001). The inverse relationship between aggregation and sliding motility is further consistent with the hypermotility of a Dlsr2 mutant on a semi-solid substratum (Arora et al., 2008). However, GPL-dependent attachment and motility appears to be unimportant for biofilm development on an air–liquid interface, likely because the liquid permits

free diffusion and growth of suspended aggregates. Biofilms formed by cells without GPLs on a solid substratum are likely to be poorly attached, but developmentally matured. Thus, GPLs could have a specialized role in formation of anchored biofilms under shear fluid flow. This process could be relevant for environmental mycobacteria like M. avium, which is abundantly found in showerheads and sewage systems (Feazel et al., 2009). Following the early events of attachment and aggregated growth ( 2 days under the studied condition), a GroEL1-dependent increase in FM synthesis is observed (Ojha et al., 2005, 2010), triggering the next phase of development. The later requirement of the chaperonin in the developmental process is also consistent with the observation that DgroEL1 strains exhibit a normal aggregation phenotype (Supporting Information Fig. S4). The sequential induction of Fe and N import genes at the day-4 and day-5 stages, respectively, indicates that there are additional discrete steps in the maturation process. The temporal nature of these steps further suggests that each step is likely characterized by a unique physiological cell state, which is determined by the requirement for specific nutrients during different phases of maturation. Although the induction of Feimport genes at the day-4 stage is consistent with a requirement for Fe at this stage, the signals leading to this requirement appear to originate prior to the GroEL1-dependent phase of the maturation process. The GroEL1-dependent induction of amt genes at the day-5 stage biofilm, but not in planktonic culture, indicates that there are biofilm-specific signals produced that are necessary for the maturation. Given that Lsr2 is required during early stages for intercellular aggregation, restoration of conjugal transfer of DNA and rifampicin tolerance in the Dlsr2 suppressor is broadly consistent with the idea that the cellular C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


phenotypes unique to biofilms emerge during the later developmental stages. Delayed and differential expression of Fe and N utilization genes could play an important role in the temporal emergence of these biofilmassociated phenotypes. In summary, the characterization of a suppressor of a biofilm defective lsr2 mutant from a secondary mutation in mps has redefined the roles of both Lsr2 and Mps in early stages of mycobacterial biofilm development, and provided an evidence for the relationship between form and functions of biofilms. In addition, the transcriptional profiling of these mutants has offered an unprecedented molecular insight into the order of genetic events in maturation of biofilms. These findings together provide a more detailed framework to further determine the biology of mycobacteria in biofilms.

Experimental procedures Bacterial strains and media A list of all strains and plasmids are in Supporting Information Tables S7 and S8. Unless indicated, M. smegmatis mc2155 was grown at 378C in 7H9ADC (Becton Dickinson) with 0.05% Tween-80 for planktonic cultures. 7H10ADC agar (Becton Dickinson) was used for plate cultures. Escherichia coli (DH5 a) was grown at 378C in LB broth or LB agar. As necessary, hygromycin, kanamycin and zeocin were added at 150, 20 or 25 lg ml21, respectively, to the growth medium. Biofilms of M. smegmatis strains were grown as described earlier (Ojha et al., 2005). Briefly, cells from 10 ll of saturated planktonic cultures were inoculated into 10 ml of antibiotic-free complete biofilm medium (modified M63 medium: 15 mM NH4SO4, 100 mM KH2PO4, 0.5% w/v casamino acids, 1.8 lM FeSO4 7H2O, 110 mM glucose, 1 mM MgSO4, 1 mM CaCl2) in either 60 mm polystyrene dishes or 12-well polystyrene plates, and incubated stationary at 308C up to 7-days. The N0 version of biofilm medium was prepared by omitting casamino acids and ammonium sulfate from the complete biofilm medium. The N1/2 version was prepared by reducing the initial concentrations of the nitrogen sources by half. Similarly, C0 and C1/2 versions of biofilm medium were prepared by either omitting glucose (for C0), or reducing it to half (for C1/2).

Screening of transposon mutant library A high-density library of 100,000 primary colonies was prepared using phMycomarT7 (Sassetti et al., 2001). The library was pooled and cultured in 7H9ADC with 0.05% w/v Tween-80 with 20 lg ml21 kanamycin until OD 1.0 (6 3 108 CFU ml21). About 100,000 cells were inoculated in 5 ml biofilm media and aspirated into a 10 ml syringe. The syringe was placed upright, with its tip sterilized and parafilm-wrapped for 5 days (Supporting Information Fig. S1). The content of the syringe was filtered through a 5 lm filter and plated on 7H10ADC with 20 lg ml21 kanamycin. Transposon insertion sites in the colonies were mapped as described earlier (Ojha et al., 2015). C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

Construction of mutants and plasmids All plasmids and oligonucleotides used in this study are listed in Supporting Information Tables S7 and S9 respectively. Mutations in M. smegmatis were constructed using previously described methods (Bardarov et al., 2002; Van Kessel et al., 2008). The isogenic deletion mutant, DMSMEG_2952, was constructed using the gene replacement technique (Bardarov et al., 2002), for which the PCR-amplified allelic exchange substrates were cloned into AflII-XbaI (left arm) and NheI-BglII (right arm) of pYUB854 respectively. The recombinant plasmid was subsequently packaged into the PacI site of phAE87. The recombinant transducing particles were transduced into mc2155. Recombinant hygr colonies were confirmed by Southern blot for the expected genotype. Recombineering was used for construction of all other mutants. For construction of mpsS11X, an unmarked DMSMEG_2952 strain containing the recombineering plasmid pJV62 was grown in 7H9 medium with 0.2% succinate and 0.05% Tween-80 at 378C until an OD600 0.5, 0.2% acetamide was added to induce the recombineering proteins. Electrocompetent cells were prepared after 4 h of induction. The 60 base-pair single stranded oligonucleotides were designed with the necessary mutations and co-electroporated with the marker plasmid pJT4. The plasmid pJT4 was constructed by cloning the hygr gene from pYUB854 into pMH94 between the XbaI and NheI sites. Electroporated colonies were plated on 7H10ADC with hygromycin (150 lg ml21) and kanamycin (20 lg ml21). Candidate revertant colonies were selected based on rough colony morphology and confirmed by sequencing. For construction of deletion strains by recombineering, allelic exchange substrates were generated by sewing-PCR on either side of a loxP flanked zeocinresistant cassette using the respective primers listed in Supporting Information Table S6. The PCR-products were electroporated into an electrocompetent recombineering strain, mc2155-pJV53-SacB, and plated on 7H10ADC with 25 lg ml21 zeocin. The genotype of zeor colonies was confirmed by PCR. The recombineering plasmid, pJV53-SacB, was removed by plating the cells on 7H10ADC with 15% sucrose, and screening sucrose resistant colonies for sensitivity to kanamycin, the marker for pJV53-SacB. The zeor marker was removed by excision using a Cre recombinase. pCre-SacB was electroporated into the mutant cells and transformants were screened for loss of zeor. The zeos colonies were screened on 7H10ADC with 15% sucrose to obtain clones without the pCre-SacB plasmid. The rescued unmarked mutants were complemented as indicated. The rescued unmarked Dlsr2 and DgroEL1 strains were used for introducing mps deletion by recombineering.

Attachment assay Briefly, cells constitutively overexpressing mCherry from an integrative plasmid were cultured in 10 ml of detergent-free 7H9ADC media until late-log phase. The cells were harvested and washed three times in 10 ml PBS, and resuspended in PBS to an OD600 of 0.1. Subsequently, a 5 ml aliquot of each sample was transferred to a polystyrene petri dish (Falcon) and incubated at 378C for 24 h. After the

806 Y. Yang et al.

䊏

incubation, liquid was aspirated from the plate and residual unattached cells were removed by three washes of 10 ml PBS, each for 5 min on a shaker at 50 rpm. Attached cells were fixed with 2% paraformaldehyde made in PBS for 15 min at room temperature. The plates were air-dried for 24 h at room temperature and examined under fluorescent microscope at 203 magnification. Micrographs were analyzed by ImageJ.

Aggregation assay Strains of M. smegmatis were inoculated from an agar plate into 10 ml of 7H9ADC media without Tween-80 and cultured until saturation. Cells were washed three times and resuspended in 10 ml of PBS. The cellular aggregates in the suspension were broken with repeated passage through an 18G needle fitted in a syringe. After vortexing for 30 s, 1 ml of the cell suspension was added to a cuvette and placed in the spectrophotometer. The initial (0 min) OD600 reading was taken the moment the lid of the spectrophotometer was closed and readings were at taken each subsequent minute for 15 min. Aggregation index scores were calculated using the following equation:

ðODT50 2 ODT5t Þ3100 Aggregation Index 5 ODT50 For microscopic analysis of the aggregates settling from suspension, cells constitutively overexpressing mCherry from an extrachromosomal plasmid were cultured and processed the same way as for the optical assay. Then 200 ml of cells, adjusted to a density of 106 cfu ml21 in detergentfree PBS, placed on a pre-focused optical dish for 10 min under inverted fluorescent microscope at 203 magnification. Appearance of aggregated at the focal plane (bottom of the dish) were recorded.

Antibiotic tolerance assay The antibiotic sensitivity of biofilms at a given time of drug exposure was determined as described earlier (Ojha et al., 2008). Briefly, two sets of each strain were cultured as biofilms in complete biofilm medium for 5 days at 308C. To one set, 400 lg ml21 rifampicin was added and the exposed biofilms were incubated at 308C for the indicated period, while the other set was incubated unexposed. At the indicated time point, the rifampicin exposed and unexposed biofilms were harvested, dispersed in PBS with 0.1% Tween-80 for 24 h, and plated on 7H10ADC to count the number of viable cells. The ratio of viable cells in exposed to unexposed biofilms was calculated to determine the percentage of surviving cells.

microgram of DNA was fragmented by incubation at 378C for 9 min in 13 reaction buffer and 25% Enzyme Mix II. The reaction was stopped by addition of stop buffer and DNA was purified using Ampure XP reagent (Agencourt). A 1:10 dilution of fragmented DNA was analyzed on the Agilent Bioanalyzer to verify that median fragment size was 350–450 bp. Sequencing adapters and Ion Xpress barcodes were ligated to fragmented DNA, nick repair was performed and DNA was purified using Ampure beads. The library was size selected for 350–400 bp fragments using an E-Gel SizeSelectTM (Life Technologies). The library was amplified via 8 cycles of PCR. Library quantitation was performed using the Ion Library Quantitation kit (Life Technologies). Following template generation by OneTouch2 and bead enrichment, the library was loaded on an Ion Torrent PGM 318 chip (Life Technologies) and the sequencing reaction was run for 850 flows for a median read length of 283 bp. Raw sequences were filtered for quality score of 20 and the sequence alignment was performed on a CLC Genomics platform.

Extraction and analysis of lipids FM analysis was performed as described previously (Ojha et al., 2008). Briefly, cells were grown as biofilms in complete biofilm medium containing 1 lCi ml21 of 14C-acetic acid, sodium salt (Perkin Elmer # NEC084H001MC). Total cells were harvested, washed and FM-rich apolar lipids were extracted in 2 ml methanol:0.3% NaCl (10:1) and 1 ml petroleum ether per 50 mg dried biomass. Apolar lipids equivalent to 10 000 cpm from the air-dried organic layer were analyzed by thin-layer chromatography in chloroform:methanol (96:4). GPL analysis was performed as described previously (Ojha et al., 2002). Briefly, indicated strains were grown in 50 ml of detergent-free 7H9ADC medium at 378C until saturation, and total lipids from harvested cells were extracted twice with 10 ml of chloroform:methanol (2:1) for 24 h at room temperature, and the extracts were pooled and dried at room temperature. The air-dried lipids were resuspended in 7 ml chloroform:methanol:water (4:2:1), and centrifuged at 2000 rpm for 10 min to collect the organic layer, which was air-dried and resuspended in 1 ml chloroform:methanol (2:1). The suspension was treated with an equal volume of 0.2 M NaOH in methanol at 378C for 30 min and then neutralized with glacial acetic acid. The organic layer was collected by centrifugation. The deacylated lipids obtained from the dried organic layer were then developed on a TLC plate with chloroform:methanol (9:1). GPL was visualized by spraying the plate with 0.2% orcinol in 10% ethanolic H2SO4, followed by 10 min of baking at 1218C.

RNA-seq Genome sequencing and analysis Genomic DNA from the described strains was isolated and a library preparation of the fragments was performed using the Ion XpressTM Plus Fragment Library kit (Life Technologies) according to manufacturer’s instructions. One

The mc2155 and its mutant derivatives were grown on 7H10ADC plates at 378C to obtain a confluent lawn. Total RNA was extracted from the lawn of cells using a Qiagen RNeasy kit and contaminating DNA was removed with the turbo DNA-free kit (Thermo Fisher Scientific). For each sample, 5 mg of total RNA was processed for rRNA removal C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


using the RiboZero kit (Illumina). Strand-specific DNA libraries were then prepared with 100 ng of mRNA using the Scriptseq Complete Kit- Bacteria (Illumina). Libraries were sequenced on the NextSeq500 platform (Illumina). Sequences were analyzed by Rockhopper (McClure et al., 2013) at default settings using the reference genome of M. smegmatis mc2155 (NC_008596). RNA-seq of planktonic and pellicle biofilms of mc2155 were performed similarly, but cells were grown in detergent-free modified M63 biofilm medium as previously described (Ojha et al., 2005).

RT-qPCR All oligonucleotides used for RT-qPCR are listed in Supporting Information Table S9. DNA-free RNA for RT-qPCR was extracted as described for RNA-seq. For each sample, 200 ng of RNA was used for reverse transcription using the Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). RT-qPCR was performed on a 7000 fast RT-qPCR System (Applied Biosystems) with SYBR Green Master Mix following the manufacturer’s instructions. Relative expression of target gene is calculated as 22DCt (gene of interest-sigA).

Filter mating assay Filter mating assay was done according to the protocol as described previously (Nguyen et al., 2010). Briefly, donor and recipient strains were cultured in tryptic soy broth (TSB) with 0.05%Tween-80 at 378C until an OD600 of 1.0. Aliquots of 0.5 ml each of donor and recipient cells were mixed together and pelleted by centrifugation. The cells were washed twice in TSB, and were resuspended in a final volume of 100 ll TSB. Cells were spotted onto a filter, which was placed on tryptic soy agar (TSA) and incubated at 308C for 24 h. The cells were removed from the filter and resuspended in 1 ml TSB with 0.05%Tween-80 by vortexing. The numbers of donors, recipients, and transconjugants were determined by plating cells onto TSA plates with appropriate antibiotics that selected for each population. The conjugation frequency was expressed as the number of transconjugants per donor.

Acknowledgements The authors acknowledge technical help from Kathleen Kulka, Mohammad Islam, Timothy Connors, Jacob Richards and Natalie Boucher, and gifts of phT7MycoMar, pJV62 and pJV53-SacB from Eric Rubin, Graham Hatfull and Todd Gray respectively. Authors acknowledge two Wadsworth Center Core facilities – Applied Genomic Technologies, and Media and Tissue Culture – for their support, as well as Dr. Kimberly Mclive Reed of Health Research Inc. for reading and editing the manuscript. This work was supported by grants from NIH to AKO (AI107595 and AI132422), WRJ (AI26170) and KMD (AI107258). Y.Y., J.T., K.M.D., W.R.J. and A.O. designed experiments. Y.Y., J.T., Y.L. and C.V. performed experiments. Y.Y., K.M.D., W.R.J., and A. O. analyzed data. Y.Y., K.M.D. and A.O. wrote the manuscript. C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

Competing interest The authors declare that they have no competing financial interests.

References Amon, J., Brau, T., Grimrath, A., Hanssler, E., Hasselt, K., Holler, M., et al. (2008) Nitrogen control in Mycobacterium smegmatis: nitrogen-dependent expression of ammonium transport and assimilation proteins depends on the OmpR-type regulator GlnR. J Bacteriol 190: 7108–7116. Anderson, G.G., and O’toole, G.A. (2008) Innate and induced resistance mechanisms of bacterial biofilms. Curr Top Microbiol Immunol 322: 85–105. Arora, K., Whiteford, D.C., Lau-Bonilla, D., Davitt, C.M., and Dahl, J.L. (2008) Inactivation of lsr2 results in a hypermotile phenotype in Mycobacterium smegmatis. J Bacteriol 190: 4291–4300. Bardarov, S., Bardarov, S., Jr., Pavelka, M.S., Jr., Sambandamurthy, V., Larsen, M., Tufariello, J., et al. (2002) Specialized transduction: an efficient method for generating marked and unmarked targeted gene disruptions in Mycobacterium tuberculosis, M. bovis BCG and M. smegmatis. Microbiology 148: 3007–3017. Billman-Jacobe, H., McConville, M.J., Haites, R.E., Kovacevic, S., and Coppel, R.L. (1999) Identification of a peptide synthetase involved in the biosynthesis of glycopeptidolipids of Mycobacterium smegmatis. Mol Microbiol 33: 1244–1253. Bosio, S., Leekha, S., Gamb, S.I., Wright, A.J., Terrell, C.L., and Miller, D.V. (2012) Mycobacterium fortuitum prosthetic valve endocarditis: a case for the pathogenetic role of biofilms. Cardiovasc Pathol 21: 361–364. Camilli, A., and Bassler, B.L. (2006) Bacterial smallmolecule signaling pathways. Science 311: 1113–1116. Chen, J.M., German, G.J., Alexander, D.C., Ren, H., Tan, T., and Liu, J. (2006) Roles of Lsr2 in colony morphology and biofilm formation of Mycobacterium smegmatis. J Bacteriol 188: 633–641. Colangeli, R., Helb, D., Vilcheze, C., Hazbon, M.H., Lee, C.G., Safi, H., et al. (2007) Transcriptional regulation of multi-drug tolerance and antibiotic-induced responses by the histone-like protein Lsr2 in M. tuberculosis. PLoS Pathog 3: e87. Conrady, D.G., Brescia, C.C., Horii, K., Weiss, A.A., Hassett, D.J., and Herr, A.B. (2008) A zinc-dependent adhesion module is responsible for intercellular adhesion in staphylococcal biofilms. Proc Natl Acad Sci U S A 105: 19456–19461. Cramton, S.E., Gerke, C., Schnell, N.F., Nichols, W.W., and Gotz, F. (1999) The intercellular adhesion (ica) locus is present in Staphylococcus aureus and is required for biofilm formation. Infect Immun 67: 5427–5433. Cucarella, C., Solano, C., Valle, J., Amorena, B., Lasa, I., and Penades, J.R. (2001) Bap, a Staphylococcus aureus surface protein involved in biofilm formation. J Bacteriol 183: 2888–2896.

808 Y. Yang et al.

䊏

Davies, D. (2003) Understanding biofilm resistance to antibacterial agents. Nat Rev Drug Discov 2: 114–122. Davies, D.G., Parsek, M.R., Pearson, J.P., Iglewski, B.H., Costerton, J.W., and Greenberg, E.P. (1998) The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 280: 295–298. Etienne, G., Villeneuve, C., Billman-Jacobe, H., AstarieDequeker, C., Dupont, M.A., and Daffe, M. (2002) The impact of the absence of glycopeptidolipids on the ultrastructure, cell surface and cell wall properties, and phagocytosis of Mycobacterium smegmatis. Microbiology 148: 3089–3100. Feazel, L.M., Baumgartner, L.K., Peterson, K.L., Frank, D.N., Harris, J.K., and Pace, N.R. (2009) Opportunistic pathogens enriched in showerhead biofilms. Proc Natl Acad Sci U S A 106: 16393–16399. Fux, C.A., Costerton, J.W., Stewart, P.S., and Stoodley, P. (2005) Survival strategies of infectious biofilms. Trends Microbiol 13: 34–40. Gordon, B.R., Li, Y., Wang, L., Sintsova, A., van Bakel, H., Tian, S., et al. (2010) Lsr2 is a nucleoid-associated protein that targets AT-rich sequences and virulence genes in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 107: 5154–5159. Hall-Stoodley, L., Brun, O.S., Polshyna, G., and Barker, L.P. (2006) Mycobacterium marinum biofilm formation reveals cording morphology. FEMS Microbiol Lett 257: 43–49. Heilmann, C., Schweitzer, O., Gerke, C., Vanittanakom, N., Mack, D., and Gotz, F. (1996) Molecular basis of intercellular adhesion in the biofilm-forming Staphylococcus epidermidis. Mol Microbiol 20: 1083–1091. Hogan, D.A., and Kolter, R. (2002) Pseudomonas-Candida interactions: an ecological role for virulence factors. Science 296: 2229–2232. Hoiby, N., Bjarnsholt, T., Givskov, M., Molin, S., and Ciofu, O. (2010) Antibiotic resistance of bacterial biofilms. Int J Antimicrob Agents 35: 322–332. Islam, M.S., Richards, J.P., and Ojha, A.K. (2012) Targeting drug tolerance in mycobacteria: a perspective from mycobacterial biofilms. Expert Rev Anti Infect Ther 10: 1055– 1066. Jenkins, V.A., Barton, G.R., Robertson, B.D., and Williams, K.J. (2013) Genome wide analysis of the complete GlnR nitrogen-response regulon in Mycobacterium smegmatis. BMC Genom 14: 301. Kocincova, D., Singh, A.K., Beretti, J.L., Ren, H., Euphrasie, D., Liu, J., et al. (2008) Spontaneous transposition of IS1096 or ISMsm3 leads to glycopeptidolipid overproduction and affects surface properties in Mycobacterium smegmatis. Tuberculosis (Edinb) 88: 390–398. Leid, J.G., Willson, C.J., Shirtliff, M.E., Hassett, D.J., Parsek, M.R., and Jeffers, A.K. (2005) The exopolysaccharide alginate protects Pseudomonas aeruginosa biofilm bacteria from IFN-gamma-mediated macrophage killing. J Immunol 175: 7512–7518. Lopez, D., Vlamakis, H., and Kolter, R. (2010) Biofilms. Cold Spring Harb Perspect Biol 2: a000398. Mah, T.F., and O’toole, G.A. (2001) Mechanisms of biofilm resistance to antimicrobial agents. Trends Microbiol 9: 34–39. Malik, A., Sakamoto, M., Hanazaki, S., Osawa, M., Suzuki, T., Tochigi, M., and Kakii, K. (2003) Coaggregation

among nonflocculating bacteria isolated from activated sludge. Appl Environ Microbiol 69: 6056–6063. Marsollier, L., Brodin, P., Jackson, M., Kordulakova, J., Tafelmeyer, P., Carbonnelle, E., et al. (2007) Impact of Mycobacterium ulcerans biofilm on transmissibility to ecological niches and Buruli ulcer pathogenesis. PLoS Pathog 3: e62. McClure, R., Balasubramanian, D., Sun, Y., Bobrovskyy, M., Sumby, P., Genco, C.A., et al. (2013) Computational analysis of bacterial RNA-Seq data. Nucleic Acids Res 41: e140. Moorthy, S., and Watnick, P.I. (2005) Identification of novel stage-specific genetic requirements through whole genome transcription profiling of Vibrio cholerae biofilm development. Mol Microbiol 57: 1623–1635. Nguyen, K.T., Piastro, K., Gray, T.A., and Derbyshire, K.M. (2010) Mycobacterial biofilms facilitate horizontal DNA transfer between strains of Mycobacterium smegmatis. J Bacteriol 192: 5134–5142. Ojha, A., and Hatfull, G.F. (2007) The role of iron in Mycobacterium smegmatis biofilm formation: the exochelin siderophore is essential in limiting iron conditions for biofilm formation but not for planktonic growth. Mol Microbiol 66: 468–483. Ojha, A.K., Varma, S., and Chatterji, D. (2002) Synthesis of an unusual polar glycopeptidolipid in glucose-limited culture of Mycobacterium smegmatis. Microbiology 148: 3039–3048. Ojha, A., Anand, M., Bhatt, A., Kremer, L., Jacobs, W.R., Jr., and Hatfull, G.F. (2005) GroEL1: a dedicated chaperone involved in mycolic acid biosynthesis during biofilm formation in mycobacteria. Cell 123: 861–873. Ojha, A.K., Baughn, A.D., Sambandan, D., Hsu, T., Trivelli, X., Guerardel, Y., et al. (2008) Growth of Mycobacterium tuberculosis biofilms containing free mycolic acids and harbouring drug-tolerant bacteria. Mol Microbiol 69: 164– 174. Ojha, A.K., Trivelli, X., Guerardel, Y., Kremer, L., and Hatfull, G.F. (2010) Enzymatic hydrolysis of trehalose dimycolate releases free mycolic acids during mycobacterial growth in biofilms. J Biol Chem 285: 17380–17389. Ojha, A.K., Hatfull, G.F., and Jacobs, W.R., Jr. (2015) Genetic dissection of mycobacterial biofilms. In Mycobacteria Protocols. Parish, T. and Roberts, D. (eds). New York: Springer. Petrova, O.E., and Sauer, K. (2009) A novel signaling network essential for regulating Pseudomonas aeruginosa biofilm development. PLoS Pathog 5: e1000668. Prabhakara, R., Harro, J.M., Leid, J.G., Harris, M., and Shirtliff, M.E. (2011) Murine immune response to a chronic Staphylococcus aureus biofilm infection. Infect Immun 79: 1789–1796. Pratt, L.A., and Kolter, R. (1998) Genetic analysis of Escherichia coli biofilm formation: roles of flagella, motility, chemotaxis and type I pili. Mol Microbiol 30: 285–293. Recht, J., and Kolter, R. (2001) Glycopeptidolipid acetylation affects sliding motility and biofilm formation in Mycobacterium smegmatis. J Bacteriol 183: 5718–5724. Sakuragi, Y., and Kolter, R. (2007) Quorum-sensing regulation of the biofilm matrix genes (pel) of Pseudomonas aeruginosa. J Bacteriol 189: 5383–5386. C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V


Sassetti, C.M., Boyd, D.H., and Rubin, E.J. (2001) Comprehensive identification of conditionally essential genes in mycobacteria. Proc Natl Acad Sci U S A 98: 12712– 12717. Siegrist, M.S., Unnikrishnan, M., McConnell, M.J., Borowsky, M., Cheng, T.Y., Siddiqi, N., et al. (2009) Mycobacterial Esx-3 is required for mycobactin-mediated iron acquisition. Proc Natl Acad Sci U S A 106: 18792–18797. Stanley, N.R., and Lazazzera, B.A. (2004) Environmental signals and regulatory pathways that influence biofilm formation. Mol Microbiol 52: 917–924. Stoodley, P., Sauer, K., Davies, D.G., and Costerton, J.W. (2002) Biofilms as complex differentiated communities. Annu Rev Microbiol 56: 187–209.

C 2017 John Wiley & Sons Ltd, Molecular Microbiology, 105, 794–809 V

Van Kessel, J.C., Marinelli, L.J., and Hatfull, G.F. (2008) Recombineering mycobacteria and their phages. Nat Rev Microbiol 6: 851–857. Yamazaki, Y., Danelishvili, L., Wu, M., Hidaka, E., Katsuyama, T., Stang, B., et al. (2006) The ability to form biofilm influences Mycobacterium avium invasion and translocation of bronchial epithelial cells. Cell Microbiol 8: 806–814.

Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web-site.