Rv0004 is a new essential member of the mycobacterial DNA replication machinery

DNA replication is fundamental for life, yet a detailed understanding of bacterial DNA replication is limited outside the organisms Escherichia coli and Bacillus subtilis. Many bacteria, including mycobacteria, encode no identified homologs of helicase loaders or regulators of the initiator protein DnaA, despite these factors being essential for DNA replication in E. coli and B. subtilis. In this study we discover that a previously uncharacterized protein, Rv0004, from the human pathogen Mycobacterium tuberculosis is essential for bacterial viability and that depletion of Rv0004 leads to a block in cell cycle progression. Using a combination of genetic and biochemical approaches, we found that Rv0004 has a role in DNA replication, interacts with DNA and the replicative helicase DnaB, and affects DnaB-DnaA complex formation. We also identify a conserved domain in Rv0004 that is predicted to structurally resemble the N-terminal protein-protein interaction domain of DnaA. Mutation of a single conserved tryptophan within Rv0004’s DnaA N-terminal-like domain leads to phenotypes similar to those observed upon Rv0004 depletion and can affect the association of Rv0004 with DnaB. In addition, using live cell imaging during depletion of Rv0004, we have uncovered a previously unappreciated role for DNA replication in coordinating mycobacterial cell division and cell size. Together, our data support that Rv0004 encodes a homolog of the recently identified DciA family of proteins found in most bacteria that lack the DnaC-DnaI helicase loaders in E. coli and B. subtilis. Therefore, the mechanisms of Rv0004 elucidated here likely apply to other DciA homologs and reveal insight into the diversity of bacterial strategies in even the most conserved biological processes.


Author summary
DNA is the molecule that encodes all of the genetic information of an organism. In order to pass genes onto the next generation, DNA has to first be copied through a process called DNA replication. Most of the initial studies on bacterial DNA replication were performed in Escherichia coli and Bacillus subtilis. While these studies were very informative, a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Introduction
The ability to maintain, replicate, and express genetic information encoded in DNA is critical to all domains of life. DNA replication studies in Escherichia coli and Bacillus subtilis have elucidated the mechanisms of bacterial DNA replication initiation, elongation, and termination, but the applicability of many of these findings to other bacteria is less clear. Briefly, initiation begins when DnaA, the initiator protein, binds to specific sites located at the origin of replication (oriC) and oligomerizes, forming a nucleoprotein complex that results in the melting of the adjacent DNA [1]. Next, helicase loaders and accessory primosomal proteins, with the help of DnaA, load the replicative helicase onto melted DNA [2,3]. The replicative helicase then binds the primase, which lays down short RNA primers [2]. Clamp loader complexes load DNA Polymerase III (Pol III) onto primed DNA, allowing replication elongation to begin [1]. Elongation proceeds bi-directionally from oriC until it reaches termination sites bound by terminator proteins [4].
Although the general stages of DNA replication are likely conserved in all bacteria, many steps have not been studied outside of the model organisms E. coli and B. subtilis. In particular, DNA replication is not well understood in mycobacteria, including the human pathogen Mycobacterium tuberculosis (Mtb). While high-fidelity DNA replication and repair is critical to maintain chromosomal integrity, mutations generated by error-prone DNA replication can enhance Mtb virulence and lead to antibiotic resistance [5]. The study of DNA replication and repair in mycobacteria is particularly relevant given that all acquired drug resistance in Mtb arises through chromosomally encoded mutations [6,7]. Mycobacteria encode homologs of some, but not all, DNA replication proteins and it is not clear how most mycobacterial homologs function relative to their E. coli counterparts. For example, Rock et al. recently showed that the DNA Pol III ε exonuclease, which is essential for replication fidelity in E. coli, is dispensable in Mtb [8]. There are also a number of processes essential for efficient DNA replication in E. coli and B. subtilis for which homologs have not been identified in mycobacteria, including regulators of DnaA activity (Hda in E. coli, YabA in B. subtilis [9]), proteins that load the replicative helicase (DnaC in E. coli, DnaI in B. subtilis [2,10]), and replication terminator proteins (Tus in E. coli and RTP in B. subtilis [11]). The proteins required for these processes in E. coli and B. subtilis are functionally analogous but are not conserved in sequence. Therefore, functionally similar mycobacterial proteins could exist and remain unidentified due to sequence divergence.
In this study we have discovered that Rv0004 in Mtb and MSMEG_0004 in Mycobacterium smegmatis are essential for DNA replication even though they are absent from E. coli and B. subtilis, the organisms traditionally used to study bacterial DNA replication. Rv0004 had never before been studied but is predicted to contain a domain of unknown function 721 (DUF721, PF05258). In a recent publication, Brezellec et al. used bioinformatics to identify DUF721-containing proteins in 23 out of 26 bacterial phyla and named the members of this protein family "DciA," for DnaC DnaI antecedent, based on the finding that these proteins preceded the DnaC-DnaI helicase loading systems of E. coli and B. subtilis [12]. While Brezellec et al. illustrate how widely distributed DciA homologs are, which underscores the importance of work on DciA proteins to the field of bacterial DNA replication, the experimental data is limited to showing that Pseudomonas aeruginosa DciA is important for DNA replication in P. aeruginosa and associates with DnaB in a bacterial two-hybrid assay [12]. Therefore, a molecular and biochemical analysis of DciA has yet to be performed and is necessary to define the basis of DciA's interaction with DnaB outside of the bacterial two-hybrid system, to determine how DciA's association with DnaB relates to its role in DNA replication, and to elucidate other activities for DciA in the cell.
In this manuscript we perform the first mechanistic studies on a DciA homolog and show that Mtb Rv0004 (DciA Mtb ) directly binds DNA and the replicative helicase DnaB to regulate the interaction of DnaB with the initiator protein DnaA. We discover that the DUF721 in DciA proteins comprises a protein-protein interaction domain that is predicted to structurally resemble the N-terminus of DnaA. We provide data to support the importance of this domain by showing that the mutation of a single conserved tryptophan within DciA Mtb 's DnaA N-terminal-like domain leads to defects in DciA Mtb 's cellular activity and can affect the association of DciA Mtb with DnaB. In addition, using live cell imaging during depletion of dciA Mtb we have uncovered a previously unappreciated role for DNA replication in the coordination of mycobacterial cell division and cell size. Together, these studies elucidate a mechanism by which DciA proteins affect DNA replication initiation, identify a function for a conserved protein domain, and provide insight into the influence of DNA replication on cell cycle in mycobacteria.

Identification of MSMEG_0004 and Rv0004
In previous work we probed the transcriptional responses of M. smegmatis, a non-pathogenic model organism for Mtb, to double-stranded DNA (dsDNA) breaks [13]. We identified MSMEG_0004 as being upregulated in response to dsDNA breaks that occurred during logarithmic (log) but not stationary phase. By measuring the expression of MSMEG_0004 during log versus stationary growth phase in the absence of induced DNA damage, we found that MSMEG_0004 is also more highly expressed in log phase in the absence of stress (Fig 1A). We observed a similar result for the Mtb homolog Rv0004 (Fig 1B). Together, these data suggest that expression of 0004 genes is important while the bacteria are actively growing and dividing.
MSMEG_0004 homologs, which are not present in eukaryotes, E. coli, or B. subtilis, encode hypothetical proteins that contain a domain of unknown function (DUF721; PFAM05258, Fig  1C) and are predicted to be nucleic acid-binding proteins (COG5512 family members) [14]. Brezellec et al. recently identified DUF721 as being widely conserved in proteins they termed DciA [12]. Based on the presence of DUF721 in Rv0004 and MSMEG_0004, we will refer to Rv0004 as dciA Mtb and MSMEG_0004 as dciA Msm .
dciA Msm is located in an operon next to oriC that also contains dnaN and recF, which encode the DNA Pol III beta clamp and a DNA repair protein, respectively ( Fig 1D) [15,16]. This operon is located between dnaA and the gyrB-gyrA operon, which encode the replication initiator protein and bacterial gyrase, respectively. This genome structure is conserved between Rv0004 is essential for DNA replication in mycobacteria M. smegmatis and Mtb except for MSMEG_0002, a gene that is predicted to encode a 6-phosphogluconate dehydrogenase and is encoded at a separate genomic location in Mtb (Fig 1D) [17]. The orientation of the DNA replication and repair genes dnaA, dnaN, recF, gyrB, gyrA near oriC is a conserved feature in many bacteria [18]. Like dciA Msm , dnaA, dnaN, and recF were all more highly expressed in log phase versus stationary phase, consistent with roles in DNA replication (S1 Fig). The genomic location and the increased expression of dciA Mtb and dciA Msm during log phase support a link between the mycobacterial DciA proteins and DNA replication or repair.
dciA is essential for M. smegmatis viability in culture In order to study the roles for dciA Msm and dciA Mtb in mycobacteria, we took a reverse genetic approach. Attempts to delete dciA Msm from M. smegmatis and dciA Mtb from Mtb were unsuccessful, suggesting these genes are essential for viability. To study the dciA genes, we constructed a merodiploid strain in which the dciA Mtb gene was integrated at the M. smegmatis attB site under the control of a promoter that contains tet operator sites and is linked to a kanamycin resistance cassette (Fig 2A, S1 and S2 Tables). We then deleted dciA Msm from its endogenous locus (Fig 2B), resulting in the strain ΔdciA Msm attB::tetdciA Mtb (Fig 2A). We used the ΔdciA Msm attB::tetdciA Mtb strain to study dciA Mtb in the context of M. smegmatis. Unfortunately, we were unable to engineer similar strains in Mtb, likely due to the difficulty of manipulating the genome near oriC. To confirm that dciA is essential in M. smegmatis, we used a gene-switching technique [19] to replace the kanamycin resistance cassette-linked dciA Mtb allele at the attB site in the ΔdciA Msm attB::tetdciA Mtb strain with a zeocin-resistant plasmid that was either empty or expressed dciA Mtb . We were able to switch in the zeocin-resistant plasmid expressing dciA Mtb but we were unable to recover a dciA gene null mutant (Fig 2C), further supporting that dciA Msm is essential for viability.

Depletion of dciA Mtb results in inhibition of growth
To study the loss of dciA expression in M. smegmatis, the ΔdciA Msm attB::tetdciA Mtb strain was transformed with an episomal plasmid expressing a Tet-ON repressor (TetR) [20] (S1 and S2 Tables). In this M. smegmatis ΔdciA Msm attB::tetdciA Mtb +pTetR strain, referred to as Tet-DciA, dciA Mtb transcript is only expressed in the presence of anhydrotetracycline (ATc). We monitored dciA Mtb depletion by diluting cultures of Tet-DciA grown in the presence of ATc (+ATc) into liquid media lacking ATc (-ATc) and collecting RNA at 3,8,16,24,28, and 32 hours after growth -ATc. After 16 hours in depleting conditions, dciA Mtb transcript levels were 96% lower than at 3 hours (Fig 2D and 2E). To characterize the requirement of dciA Mtb expression for growth, we monitored the growth of the Tet-DciA strain during dciA Mtb transcript depletion. Over the first 24 hours, there was no significant difference in the growth of Tet-DciA in depleting (-ATc) or replete (+ATc) conditions ( Fig 2F). However, after diluting the stationary phase culture back to early log phase at 24 hours, depleted cells grew more slowly than controls (Fig 2F). We calculated the doubling times of Tet-DciA cultured ± ATc and found that after 24 hours, M. smegmatis depleted of dciA Mtb grew significantly more slowly than the replete controls (5.8 hour versus 3.7 hour doubling time) (Fig 2G).
To determine whether the growth defect of dciA Mtb -depleted cells was due to an inability to recover from stationary phase versus DciA Mtb playing a critical role in growth during log phase, we performed a continual log liquid growth curve where cultures were diluted to early log phase every 8 hours to ensure that they did not enter stationary phase. Similar to the standard growth curve experiments, M. smegmatis depleted of dciA Mtb did not show a significant growth defect until after 24 hours of growth-ATc ( Fig 2H). We calculated the doubling times of Tet-DciA cultured ± ATc and found that after 24 hours of continual log growth, M. smegmatis depleted of dciA Mtb grew significantly more slowly than the replete controls (4.7 hour versus 3.4 hour doubling time) (Fig 2I). After 40 hours of growth in depleting conditions, suppressors of the Tet-ON system are selected for and the levels of dciA Mtb transcript are no longer controlled by ATc. Together, these data demonstrate that DciA is important for M. smegmatis growth.

Depletion of dciA Mtb results in a block in cell cycle progression
We have shown that dciA is essential for growth in culture and is involved in responding to DNA damage in M. smegmatis. However, a role in DNA damage responses alone cannot explain the essentiality of dciA, since genes key for mycobacterial DNA repair pathways are not essential in vitro [21][22][23][24]. The location of mycobacterial dciA genes in an operon with and adjacent to genes involved in DNA replication raises the question of whether DciA plays a role in this essential process in mycobacteria. In addition, P. aeruginosa DciA (DciA Pa ) has been implicated in DNA replication [12]. To investigate a role for mycobacterial DciA in DNA replication and the bacterial cell cycle, we monitored the cellular morphology of Tet-DciA cultured ± ATc in the continual log growth curve ( Fig 2H). By 24 hours of depletion, Tet-DciA cultured without ATc were on average 12.9% (p <0.01) longer than Tet-DciA cultured in To characterize where in the cell cycle dciA Mtb -depleted M. smegmatis is blocked, we analyzed nucleoid morphology and septum formation. To observe nucleoid morphology, Tet-DciA was grown ± ATc, DNA was stained with DAPI, and cells were visualized by fluorescent microscopy (Fig 3A). The nucleoid in dciA Mtb -replete Tet-DciA (+ATc) appears as several distinct puncta throughout the length of the cell (Fig 3A, top row, second panel), as has been reported previously for M. smegmatis [25]. Following dciA Mtb depletion, the DAPI-stained nucleoid still appeared as several distinct puncta, but was not distributed throughout the length of the cell. Instead, there were areas free of DNA staining found at the poles ( Fig 3A). We quantified the abnormal nucleoid morphology using "% DNA occupation," which represents the nucleoid length as a percentage of the total cell length. Tet-DciA depleted for dciA Mtb exhibit significantly lower % DNA occupation starting at 16 hours of depletion (Fig 3C and  3D; S2 Fig). The observation that depletion of dciA Mtb leads to significantly lower % DNA occupation at 16 hours even though cell lengths are not significantly longer at this time point (S2 Fig) indicates that abnormal nucleoid morphology is the earlier phenotype, and that a DNA-related function causes the cell cycle block and slowed growth (Fig 2I).
We also observed the presence of 9% anucleate cells by DAPI staining upon dciA Mtb depletion ( Fig 3A). Anucleate cells result when DNA replication and cellular division are uncoupled. The presence of anucleate cells also indicates that cell division is still able to occur. To confirm that septum formation was intact, we used transmission electron microscopy (TEM) and found that cells depleted for dciA Mtb were still able to form normal septa (Fig 3E).
To visualize cell division of dciA Mtb -depleted M. smegmatis in real time, we performed livecell imaging [26] of Tet-DciA grown ± ATc in a constant-flow microfluidic device (S1 and S2 Movies). Using FM4-64 membrane stain, we observed that dciA Mtb -depleted cells can form Rv0004 is essential for DNA replication in mycobacteria septa and undergo cell division (S2 Movie). We also detected increased cell death in dciA Mtbdepleted cells compared to controls, where dead cells stop elongating and take up more FM4-64 dye, thus exhibiting a rapid increase in fluorescence (S2 Movie). 9.4% of dciA Mtb -depleted cells displayed these cell death characteristics. In support of our fixed fluorescent microscopy, Tet-DciA cells grown -ATc were on average 56% longer at birth and division than Tet-DciA grown +ATc (S3 Fig, S1 and S2 Movies, p <0.0001). However, despite the increased average length of dciA Mtb -depleted cells, we also observed division of unusually small cells (S3E Fig), many of which died shortly after division. In general, we observed greater variability in birth length in dciA Mtb depleted versus replete cells, with coefficients of variation of 41.7% and 18.2%, respectively (S3C Fig). The increased heterogeneity among cell birth lengths during dciA Mtb depletion indicates a disruption in the coordination of septum formation with cell growth, leading to the dysregulation of cell size. Thus, time-lapse imaging demonstrates that in addition to dciA Mtb -depleted cells being elongated on average, they are also characterized by increased variation in cell size and frequency of death.
Together, our data show that cell division is intact but chromosome replication or segregation is blocked during dciA Mtb depletion. In support of this conclusion, the abnormal nucleoid morphologies observed in M. smegmatis depleted of dciA Mtb phenocopy those of M. smegmatis depleted of DnaA, the chromosomal replication initiator protein [25] (

dciA depletion results in decreased DNA synthesis
The data so far support a model that DciA homologs are required for either DNA replication or chromosome segregation. To determine if dciA Mtb -depleted M. smegmatis is defective in DNA replication, we directly measured rates of DNA synthesis using a nucleotide incorporation assay [25,28]. Specifically, we determined the rates of [5,6-3 H]-thymidine incorporation into DNA by Tet-DciA cells grown ± ATc in continual log phase. We found that the rate of DNA synthesis in M. smegmatis was significantly lower at 24 hours (Fig 5A and 5C) and 36 hours (Fig 5B and 5C) in Tet-DciA grown -ATc relative to Tet-DciA grown +ATc, proving that the rate of DNA replication itself decreases upon dciA Mtb depletion. To further confirm this defect in DNA replication, we stained Tet-DciA grown ± ATc with DAPI and measured DNA content using flow cytometry. dciA Mtb -depleted cells had lower DNA content per cell based on the mean fluorescence intensity (MFI) of DAPI staining relative to controls ( Fig 5D  and S6A Fig). Together, these data demonstrate that DciA Mtb is involved in DNA replication, but do not differentiate between which step(s) of DNA replication DciA Mtb acts.

DciA Mtb binds DNA
To determine how DciA proteins function in DNA replication, we investigated macromolecular interaction partners of DciA Mtb . DciA Mtb has a calculated isoelectric point (pI) around 12, indicating that the protein is positively-charged at neutral pH. Other proteins in mycobacteria with high isoelectric points include histone-like proteins (H-NS, HupB) and integration host factor (IHF), which are all known to bind DNA [29][30][31][32]. DciA Mtb , as a member of COG5512, is also predicted to be a nucleic-acid binding protein [14]. Binding to nucleic acid could be relevant to the role for DciA in DNA replication given the numerous protein-nucleic acid complexes that form during this process. We tested the DNA binding activity of purified DciA Mtb protein (S7A Fig) in electromobility shift assays (EMSAs) with 32 P-radiolabeled DNA. Due to its role in DNA replication (Figs 3 and 5), we first tested whether DciA Mtb was able to bind oriC DNA in vitro. DciA Mtb was able to bind and shift a 553 basepair (bp) dsDNA fragment containing Mtb oriC DNA [33] (Fig 6A). Given its high isoelectric point and the negative charge of DNA, we hypothesized that DciA Mtb would be able to bind any DNA sequence and not just oriC. Indeed, we found that DciA Mtb was also able to shift a 333 bp dsDNA sequence from a site in the genome distantly located from oriC (S7B Fig). DciA Mtb 's DNA binding activity is not limited to dsDNA sequences as DciA Mtb was also able to bind and shift a 72 nucleotide single-stranded DNA (ssDNA) oligo (S7C Fig). These data indicate that DciA Mtb can bind diverse double and single stranded DNA molecules in vitro, including oriC. This sequence independent DNA-binding activity could relate to the role of DciA Mtb in DNA replication. strain, in the presence (top row) or absence (bottom two rows) of ATc pictured in brightfield, stained with DAPI for DNA, and stained with FM1-43FX for membrane. Scale bars are 10 μm. (F) Cell length and (G) % DNA occupation histograms of csm362 cells after 24 hours of growth in depleting (-ATc) or replete (+ATc) conditions. (H)   We next identified the mycobacterial proteins that associate with DciA Mtb by performing coimmunoprecipitation mass spectrometry (co-IP/MS). We engineered a strain of M. smegmatis that encodes an HA-tagged version of dciA Mtb (HA-DciA Mtb ) as its only dciA allele (S1 Table). We generated cell lysate from this strain and immunoprecipitated HA-DciA Mtb along with associated protein complexes using an anti-HA antibody conjugated agarose (Sigma). After eluting with HA peptide, eluates were separated by SDS-PAGE, silver-stained, and bands spe-  Table).

DciA Mtb interacts directly with DnaB
Since we have shown that DciA Mtb is involved in DNA replication (Fig 5), we sought to confirm the association of DciA Mtb with the co-immunoprecipitated DNA replication proteins. These proteins included gyrase, DnaX (τ clamp-loader subunit of DNA Pol III), DnaB (replicative helicase), and DnaA (replication initiator protein). To prioritize our studies, we identified DNA replication proteins that are conserved in E. coli and B. subtilis but do not yet have known homologs in mycobacteria, namely DnaA regulators and DnaB helicase loaders (Table 1). Since DnaA and DnaB were both found to associate with DciA Mtb through co-IP/ MS, we first tested whether DciA Mtb directly interacts with these proteins.
We performed pull-down experiments similar to the co-IP approach described earlier, but using purified recombinant HA-DciA Mtb , DnaB, and DnaA. We immobilized HA-DciA Mtb onto anti-HA agarose and added DnaA or DnaB. Analysis of the protein complexes eluted with HA peptide showed that HA-DciA Mtb pulls down both DnaA (Fig 6C lane 4) and DnaB (Fig 6C lane 6), but not a negative control protein, Rel DnaA, and DnaB can all bind DNA, we tested whether the interactions between these proteins were dependent on nucleic acid by performing the same pull-down experiments using recombinant proteins purified from Benzonase-treated lysates to degrade nucleic acids. When proteins purified from Benzonase-treated lysates were used, HA-DciA Mtb failed to pull down DnaA (Fig 6C lane 5) but retained its interaction with DnaB (Fig 6C lane 7). Therefore, DciA Mtb interacts directly with DnaB but depends on nucleic acid to associate with DnaA. We confirmed that DciA Mtb and DnaB directly interact by performing the reciprocal pulldown with purified FLAG-tagged DnaB (DnaB-FLAG) immobilized as bait and DciA Mtb as prey (Fig 6D). We performed Bio-Layer Interferometry (BLI) to quantify the affinity of the interaction between DnaB and DciA Mtb . The association and dissociation of varying concentrations of DciA Mtb to biotinylated DnaB loaded onto streptavidin-coated biosensor pins (ForteBio) were measured ( Fig 6E) and an affinity constant (K D ) of 210.9 ± 8.19 nM was represent each replicate, center values and error bars represent mean ± SEM. (I) Fold enrichment in levels of DNA fragments containing oriC (oriC1-3, see S8A Fig), the promoter of the rplN operon (rplN), or the sigA coding sequence (sigAIN) in samples immunoprecipitated with an anti-HA antibody (αHA-IP) or input samples (input) from strains expressing untagged DciA Mtb (No tag, black bars) and HA-DciA Mtb (grey bars) strains. Data is represented as fold enrichment relative to No tag levels. Bars represent mean ± SEM (n = 5 except No tag sigAIN, which is n = 4). **** p <0.0001, * p <0.05, ns is not significant, statistical significance was determined by one-way ANOVA and Tukey's multiple comparison test.

DciA Mtb can affect the interaction between DnaB and DnaA
In E. coli, the interaction between DnaA and DnaB is required for efficient loading of DnaB [51,52]. The result that DciA Mtb directly binds DnaB and can indirectly associate with DnaA led us to probe whether DciA Mtb affects DnaB-DnaA complex formation. Mtb DnaA has been shown to interact with the N-terminus of DnaB (residues 1-206) [36], and we confirmed that the full length Mtb DnaB-FLAG and DnaA-HA proteins directly interact in our system (Fig 6F lane 1 and Fig 6G lane 6). To determine the effect of DciA Mtb on the DnaA-DnaB interaction, we performed pull-downs with DnaA-HA as bait and DnaB-FLAG as prey in the presence of varying amounts of DciA Mtb . All pull-downs were performed using proteins purified from Benzonase-treated lysates to exclude contributions from nucleic acid interactions. As increasing amounts of DciA Mtb were added, DnaA-HA pulled down more DnaB-FLAG (Fig 6F and 6H). These results indicate that DciA Mtb facilitates DnaA-DnaB complex formation. In contrast, when DnaB-FLAG was immobilized as bait and DnaA-HA was added as prey, increasing the amount of DciA Mtb present did not change the amount of DnaB-FLAG that associated with DnaA-HA (Fig 6G and 6H). Therefore, DciA Mtb can affect DnaB-DnaA complex formation, but this is dependent on which protein is immobilized. One possible explanation for this observation is that DnaB, which functions as a hexamer [34], is unable to properly hexamerize while immobilized as bait, affecting DciA Mtb activity. The ability of DciA Mtb to positively affect the DnaB-DnaA interaction suggests that DciA Mtb promotes rather than inhibits DNA replication, which is consistent with the observations that dciA Mtb depletion leads to decreased DNA synthesis and DNA content (Fig 5).

DciA Mtb is enriched at oriC in M. smegmatis
The ability of DciA Mtb to affect the association of DnaB with DnaA suggests that DciA functions during DNA replication initiation. To test whether DciA is enriched at oriC where initiation occurs, we performed chromatin immunoprecipitation quantitative PCR (ChIP-qPCR) experiments. These experiments were performed with log-phase cultures of M. smegmatis strains that express HA-tagged DciA (HA-DciA), untagged DciA (no tag), and HA-tagged CarD (HA-CarD), a mycobacterial transcription factor that associates with RNA Polymerase and is known to localize to every promoter throughout the M. smegmatis genome [53]. Protein-nucleic acid complexes were immunoprecipitated from each culture using anti-HA resin and co-immunoprecipitated DNA was probed for sequences specific for oriC (S8A Fig), the rplN promoter, and internal to the sigA gene using qPCR. Enrichment of sequences within a given sample was determined relative to the no tag control. As expected, only DNA fragments containing promoters (oriC and rplN promoter) were specifically and significantly enriched for following immunoprecipitation with HA-CarD (S8B Fig). The only DNA fragments that were significantly enriched for following immunoprecipitation of HA-DciA Mtb were those containing the oriC (Fig 6I). As a control, no sequences were enriched in input samples before immunoprecipitation (Fig 6I and S8C Fig). The specific enrichment of DciA Mtb at oriC and not at other areas of the chromosome indicates that DciA Mtb is involved in DNA replication initiation, which is consistent with its role in affecting the interaction between DnaB and the DnaA initiator protein (Fig 6F-6H).

The C-terminus of DciA Mtb is similar to the DnaA N-terminal domain
To investigate how DciA Mtb facilitates the association between DnaA and DnaB, we used structural prediction tools to gain further insight into the protein architecture. Both Phyre2 [54] and I-TASSER [55] predicted that a region in the C-terminus of DciA Mtb within DUF721 (~92-142 aa) is structurally similar to the N-terminal domain (NTD) of DnaA in B. subtilis (DnaA Bs , PDB: 4TPSD) and E. coli (DnaA Ec , PDB: 2E0GA) (Fig 7A and 7B). We named this region of DciA Mtb the DnaA NTD-Like (DANL) domain. The DnaA NTD is responsible for many protein-protein interactions important for DNA replication, including the interaction of DnaA Ec with DnaB, DiaA, and other DnaA Ec monomers, the interaction of DnaA Bs with SirA, and the interaction between H. pylori DnaA and HobA [51,52,[56][57][58]. Phenylalanine 46 (F46) in DnaA Ec , which is equivalent to F49 in DnaA Bs , is specifically important for DnaA Ec to load DnaB [57]. Though shifted by one residue, our structural alignment shows that a tryptophan at position 113 in DciA Mtb (W113) is the closest aromatic amino acid to F49 in DnaA Bs ( Fig  7A, 7C and 7D). Consurf alignment [59] reveals that the region around W113 is one of two highly conserved regions of the DciA Mtb protein that are both located in DciA Mtb 's C-terminus Highly conserved residues are in magenta and variable residues are teal. The W113 residue is underlined. Due to a probable misannotation of the start codon, the beginning of actinobacterial DciA proteins share poor homology (Fig 1C). Therefore, Consurf analysis of DciA begins at the glycine that marks the beginning of homology across mycobacterial DciA proteins (at residue 21 in Fig 1C).
( Fig 7E). The position of W113 in the predicted structural model and the conservation of this region across DciA homologs suggest that W113 may be important for DciA Mtb activity.

A single point mutation in the DciA Mtb DANL domain results in similar phenotypes as dciA depletion
To test if W113 in the DciA Mtb DANL domain is important for DciA Mtb function, we engineered M. smegmatis to expresses a version of DciA Mtb where the W113 is mutated to an alanine (W113A) as its only allele of dciA. The W113A mutation leads to a growth defect (Fig 8A  and 8B), elongated cellular and abnormal nucleoid morphologies (Fig 8C-8F), and decreased DNA content (Fig 8G; S6B Fig). The observation that the mutation of a single residue can cause similar phenotypes to those observed during dciA Mtb depletion confirms that the dciA Mtb -depletion phenotypes were not due solely to the depletion of an essential protein.
These experiments also show that the W113 residue, which is located within the region of DciA Mtb that is predicted to be structurally similar to the protein-protein interaction domain of DnaA, is important for DciA Mtb activity.

The W113A mutation affects the binding of the DciA Mtb C-terminus to DnaB
In order to determine how W113 is contributing to DciA Mtb function, we purified DciA Mtb W113A protein and tested its ability to perform the functions we have assigned to Given the location of W113 in the predicted DANL protein-protein interaction domain, we were surprised that the W113A mutation did not affect the interaction between DciA Mtb and DnaB. To test whether the DANL domain itself was involved in the interaction with DnaB, we purified the HA-tagged DciA Mtb N-terminus (HA-DciA Mtb

Δ1-88
, containing the DANL domain) protein truncations, and tested their ability to bind DnaB. We found that both the N-terminus (HA-DciA Mtb

Δ1-88
) of DciA Mtb can individually bind DnaB-FLAG (Fig 8H). While we predicted the DANL-domain within the C-terminus of DciA Mtb would bind DnaB based on its similarity to the DnaA NTD, we did not anticipate that the DciA Mtb N-terminus would also interact with DnaB. This is particularly interesting since there are no known or predicted structures for the N-terminus of DciA Mtb . We were unable to isolate viable M. smegmatis dciA Mtb Δ89-167 or dciA Mtb Δ1-88 mutants using the gene-switching approach described earlier, indicating that both halves of the protein contribute to DciA Mtb 's essential cellular function. Therefore, the ability of both the N-and C-terminus of DciA to associate with DnaB is unlikely due to redundant roles for these two regions of the protein.
We tested whether the DnaB-binding activity provided by the N-terminus of DciA Mtb was precluding our ability to assess the contribution of W113 to the interaction with DnaB. To investigate this, we purified the HA-tagged DciA Mtb W113A C-terminus (HA-DciA Mtb W113AΔ1-88 ) and tested its ability to bind DnaB-FLAG. We found that HA-DciA Mtb W113AΔ1-88 associated with more DnaB-FLAG than the wild-type HA-DciA Mtb Δ1-88 (Fig 8I). Therefore, the W113A mutation in DciA Mtb affects the association of the C-terminal DANL domain with DnaB and it Rv0004 is essential for DNA replication in mycobacteria

Discussion
DciA proteins were recently discovered in an evolutionary and phylogenetic analysis that defined them by the presence of the DUF721/PF05258 domain [12]. Two classes of dciA genes were identified, those located in the dnaN-recF operon and those located elsewhere. Using P. aeruginosa dciA (dciA Pa ) as an example of the second class of dciA genes, Brezellec et al. showed that DciA Pa can associate with DnaB in a bacterial two-hybrid assay and that DNA replication is blocked during DciA Pa depletion in P. aeruginosa [12]. However, the molecular details of how DciA associates with DnaB, the mechanisms by which DciA facilitates replication, whether DciA performs other activities in the cell, and whether the findings for DciA Pa hold true for the other class of DciA proteins remained unknown.
In this study we address these gaps in knowledge beginning with our discovery that DciA Mtb , a member of the first class of DciA proteins, is an essential component of the DNA replication machinery that directly interacts with the DnaB helicase. Through detailed mechanistic work, we then expand on the work in P. aeruginosa to show that DciA localizes to the oriC, directly binds DNA, and affects the association of DnaB with DnaA, likely contributing to a role during DNA replication initiation. We also assign a role to DUF721 as a DnaA-NTDlike (DANL) protein-protein interaction domain and identify a tryptophan within the DANL domain that is critical for DciA Mtb function in vivo and can affect DnaB binding in vitro. We find that both the N-terminus and C-terminus of DciA Mtb are able to directly bind DnaB, indicating a complicated association between DciA and the replication machinery that will be the focus of future studies. The DANL domain of DciA Mtb adds to the examples of DNA replication proteins across bacteria that share structural similarity despite divergent sequences (e.g. HobA and DiaA [58] as well as winged-helix domains of bacterial, mammalian, and archaeal DNA replication proteins [60,61]). It also remains possible that the DciA DANL domain interacts with additional proteins beyond DnaB, which will be explored in future studies.
To study the role of DciA in mycobacteria, we used a Tet-DciA strain that depletes dciA Mtb transcripts following the removal of ATc from the media. dciA Mtb transcripts were depleted by 16 hours following the removal of ATc, at which point we observed abnormal nucleoid morphology. The abnormal nucleoid morphology is the earliest phenotype observed during dciA Mtb depletion and is followed by elongated cell lengths at 24 hours and then slower growth between 24 and 36 hours of dciA Mtb depletion (S2 Fig). These data suggest that the defect in DNA replication that leads to abnormal nucleoid morphology results in the cell cycle block and a subsequent growth defect. Live-cell imaging revealed that both shorter and longer dciA Mtb -depleted cells underwent division, leading to increased heterogeneity among cell birth lengths and indicating a dysregulation of cell size and division (S1 and S2 Movies, S3 Fig). It is unknown what mechanisms control cell size in mycobacteria but our data implicates DciA in affecting the coordination of cell division and cell size. Since this is the first time a mycobacterial DNA replication mutant has been analyzed by live-cell imaging, future studies will determine if this dysregulation of cell size control is a general feature of mycobacterial DNA replication mutants or if it is DciA-specific. These studies will shed light onto the mechanisms that control mycobacterial cell size and cell cycle.
Helicase loaders have not been identified in the majority of bacterial phyla, making it tempting to assign this role to DciA. However, DciA homologs do not encode ATPase domains and, thus, do not fit the definition of a helicase loader [2]. Their small size, association with the replicative helicase, and importance in DNA replication are more reminiscent of the primosomal proteins B. subtilis DnaD and E. coli DnaT [18,62]. However, DnaD is present in the cell at very high levels [18], and neither DnaD nor DnaT has a high isoelectric point or a structurally predicted DANL domain. Therefore, DciA is unique from any protein that has been shown to facilitate DnaB association with replication complexes. There are two possible explanations for the lack of identified helicase loaders in mycobacteria. DnaB could be loaded by DciA Mtb through a novel mechanism that is independent of any external ATPase activity and thereby replaces the need for a canonical helicase loader. Alternatively, DciA Mtb could function in concert with an unknown ATPase in order to load the helicase. Our co-IP experiments identified a few known and hypothetical ATPase and ATP-binding domain containing proteins (S4 Table). Future studies will investigate the interactions of DciA Mtb with these proteins and their roles in replication, as well as whether DciA Mtb affects activities of DnaB outside of its interaction with DnaA (Fig 6F and 6H).
Our studies also revealed that DciA Mtb can interact with single and double-stranded DNA in vitro. DciA homologs have high isoelectric points that likely mediate DNA-binding ( Fig  6A). Integration host factor (IHF) also has a high isoelectric point and has been shown in E. coli to bind and bend oriC DNA [63], possibly to bring DnaA-boxes closer together and promote replication initiation [64]. The effect of IHF on oriC in mycobacteria has not yet been studied, but Lsr2, which has a high isoelectric point and is the mycobacterial functional analog of H-NS, is able to bind to open reading frames proximal to oriC [50]. ChIP-qPCR experiments showed that despite the ability of DciA Mtb to bind DNA without sequence specificity, it is enriched at the oriC. This suggests that the specific localization of DciA Mtb to oriC relies on its association with DnaB and the DNA replication initiation complex.
We originally identified DciA Msm as being upregulated in response to DNA damage [13]. Although we do not know the exact role for DciA in the response to DNA damage, one can imagine a number of possibilities. First, DNA damage and DNA replication are inextricably linked, as DNA damage can disrupt the movement of the replisome and the repair of DNA damage involves proteins that are also involved in DNA replication. Second, DciA Mtb could have a role in replication restart, which involves the loading of the replicative helicase onto non-oriC DNA and is dependent on proteins that bind DNA in a structure-dependent but sequence-independent manner in other bacteria [65]. Replication restart has not yet been studied in mycobacteria (Table 1). Lastly, DciA Mtb could have a direct role in the repair of DNA damage independent of its role in DNA replication by interacting with DNA damage repair helicases. DciA Mtb co-immunoprecipitated with a number of helicases involved in DNA repair (Fig 6B), and future studies determining whether DciA Mtb directly interacts with these helicases will help to elucidate the connection between DciA and DNA damage responses.
Although we can detect immunoprecipitated HA-DciA Mtb by MS, we have been unable to detect DciA Msm , DciA Mtb , or HA-DciA Mtb from cell lysate by western blot using a polyclonal antibody that we raised against DciA Mtb or an anti-HA antibody. Since ClpX, as well as ClpP1 and ClpC1, co-immunoprecipitated with HA-DciA Mtb (S4 Table), one explanation for the low level of DciA Mtb protein in the cell is that DciA Mtb is a target of Clp protease, which is essential in mycobacteria [66]. ClpX and ClpC1 are adaptors for ClpP, which consists of ClpP1 and ClpP2 in mycobacteria. Raju et al. used clpP1P2 depletion strains in M. smegmatis and Mtb to identify proteins that were present at higher levels during clpP1P2 depletion, suggesting that they are targets of Clp [66]. DciA Mtb protein levels were 2.53 fold higher (P value of .00285) during clpP1P2 depletion in Mtb, further supporting that DciA Mtb is targeted by Clp. Interestingly, DnaB was also found to be significantly enriched during clpP1P2 depletion in M. smegmatis and Mtb. The ClpP protease is known to regulate cell cycle progression proteins in Caulobacter crescentus, where, like mycobacteria, ClpP is essential [67]. Therefore, it is possible that regulation of DNA replication proteins comprises one of the essential roles for Clp in mycobacteria.
Together, this study has elucidated functions of DciA Mtb that are likely applicable to other DciA homologs. dciA genes are not found in E. coli and B. subtilis, the organisms that have traditionally been used to study bacterial replication, and as a result have remained undiscovered until recently, despite being widely conserved throughout the bacterial kingdom. Therefore, this study contributes to a developing new paradigm of bacterial DNA replication.

Fluorescence microscopy
M. smegmatis was collected, washed once with equal volume Phosphate Buffered Saline (PBS), resuspended in equal volume 1 μg/ml FM1-43FX (Thermo Fisher) diluted in PBS and incubated for 20 minutes at 37˚C shaking. Cells were then fixed with 3% paraformaldehyde in PBS for 30 minutes shaking at 37˚C. Fixed cells were applied to 0.1% poly-L-lysine (Sigma) treated multitest slides (MP Biomedicals) and then washed once with PBS. Cells were permeabilized by treatment with 2 mg/ml lysozyme (Sigma) at 37˚C for 30 minutes followed by 0.1% Triton-X 100 (Sigma) for exactly five minutes at room temperature [69]. Cells were rinsed with PBS and stained with DAPI (Thermo Fisher) diluted to 5μg/ml with Slow Fade Antifade Equilibration Buffer (ThermoFisher) and mounted using the Slow Fade Antifade Kit according to the manufacturer's instructions. Slides were visualized with a Zeiss Axioskop 2 Mot Plus equipped with an Axiocam MRm monochrome camera and a 100X, 1.4 numerical aperature Zeiss Plan Apochromat oil objective and images were acquired using Axiovision 4.6 software, or using an upright Zeiss Axio Imager M2 fluorescence microscope and the Zen Blue image acquisition software.

Live-cell imaging
M. smegmatis cells were grown to log phase overnight with shaking at 37˚C. M. smegmatis cells were filtered through a 10 μm filter to remove clumps before being loaded into a custom polydimethylsiloxane (PDMS) microfluidic device, as before [26]. The viewing device incorporates a main microfluidic channel for continuous flow for growth media, with a height of approximately 10-17 μm, and viewing chambers with a diameter of 60 μm and a height of 0.8−0.9 μm. 2% DMSO and 0.0625 mg/ml FM4-64 are present in the flowed medium, to stain septal membranes.
The microfluidics device was placed on an automated microscope stage housed within an environmental chamber maintained at 37˚C. M. smegmatis cells were imaged for up to 40 hours using a widefield DeltaVision PersonalDV (Applied Precision, Inc) with a hardwarebased autofocus. Cells were illuminated with an InsightSSI Solid State Illumination system every 15 minutes: FM4-64 was visualized with 475nm excitation and 679 nm emission wavelengths, and cells were imaged using transmitted light brightfield imaging. Each set of images was illuminated with identical imaging conditions that were optimized to decrease phototoxicity inherent in long-term fluorescent imaging of live cells.
Tet-DciA was grown in the presence of hygromycin and ATc overnight and for the duration of the DciA Mtb replete control movies (S1 Movie). Tet-DciA was grown in the presence of hygromycin and ATc overnight followed by growth in the presence of hygromycin alone for 10.5 hours prior to imaging and for the duration of imaging (S2 Movie). For analysis, all Tet-DciA cells grown in depleting conditions were born at least 24 hours after ATc was removed from the media.
Transmission electron microscopy with thin-sectioning M. smegmatis was fixed in 2% paraformaldehyde/2.5% glutaraldehyde (Polysciences Inc.) in 100 mM sodium cocadylate buffer, pH 7.2 for 1 hour at room temperature. Samples were washed in sodium cacodylate buffer and postfixed in 1% osmium tetroxide (Polysciences Inc.) for 1 hr. Samples were then rinsed extensively in dH 2 0 prior to en bloc staining with 1% aqueous uranyl acetate (Ted Pella Inc.) for 1 hr. Following several rinses in dH 2 0, samples were dehydrated in a graded series of ethanol and embedded in Eponate 12 resin (Ted Pella Inc.). Sections of 95 nm were cut with a Leica Ultracut UCT ultramicrotome (Leica Microsystems Inc.), stained with uranyl acetate and lead citrate, and viewed on a JEOL 1200 EX transmission electron microscope (JEOL USA Inc.) equipped with an AMT 8 megapixel digital camera and AMT Image Capture Engine V602 software (Advanced Microscopy Techniques).
After 20 and 60 minutes of labeling, 3 ml aliquots of cells were taken for processing to measure counts per minute (cpm) of 3 H incorporated into DNA or for enumeration of bacterial counts by colony forming units (cfu). A 7H9 media plus [5, H]-uracil sample was also collected. Aliquots for 3 H cpm were treated with 0.3M KOH to hydrolyze RNA and incubated at 37˚C for 24 hours. Macromolecules were then precipitated with ice cold 10% trichloroacetic acid (Sigma) and filtered onto glass 25 mm GF/C filters (GE Healthcare). Filters were dried under a heat lamp and submerged in 15 ml of Ultima Gold liquid scintillation fluid (Perkin Elmer). 3 H counts per minute (cpm) were measured using a Beckman LS6000IC scintillation counter that was programmed to read each sample for 5 minutes. All presented data indicates the cpm of that sample minus the cpm from the processed media + [5,6-3 H]-uracil control to subtract background, relative to the cfu of that sample. The efficiency of RNA hydrolysis by KOH treatment was confirmed by including a KOH-untreated control.

Measuring DNA content by flow cytometry
Cells were fixed and permeabilized as described for fluorescent microscopy [69] with an additional RNaseI f treatment (NEB), and stained with 100 μM DAPI diluted in water for 15 minutes at room temperature. Cells were then resuspended in PBS, sonicated, and passed through a 30μm filter to remove clumps before flow cytometry analysis. Samples were analyzed using a FACSAria (Becton Dickinson) and data were processed with FlowJo (Treestar). All cells (events) were included in the analysis of DAPI intensity. , dnaA, and HA-dnaA were cloned into pGEX-6P (GE Healthcare Life Sciences, S2 and S3 Tables). The plasmids were transformed into BL21(DE3) (Novagen). Mtb DnaB and DnaB-FLAG were amplified from Mtb genomic DNA following the published cloning scheme [35] to exclude the native intein, cloned into pET SUMO (Invitrogen) and transformed into BL21(DE3) cells (Novagen). Rel  was cloned into pET SUMO [70].

Immunoprecipitation and pull-down experiments
For immunoprecipitation from M. smegmatis cell lysates, 1 liter cultures of mid-logarithmic phase M. smegmatis were washed with PBS and frozen at -80˚C until processing. Cell pellets were resuspended in 10mls NP-40 Buffer (10 mM sodium phosphate, pH8, 150 mM NaCl, 1% Nonidet P-40, and Roche EDTA-free Complete protease inhibitor cocktail) and lysed using a Constant Systems Cell Disruptor (4 passes at 40k psi), and spun at 50,000g for 30 minutes to generate lysate. When indicated, some lysates were treated with DNase I (NEB) according to manufacturers instructions. 1 ml of lysate was added to 50 μL monoclonal anti-HA agarose (Sigma) and rotated at 4˚C overnight. The anti-HA agarose was then washed 3 times with NP-40 buffer and immune complexes were eluted with 500 μg/ml HA peptide (Roche) in IP elution buffer (50 mM Tris-HCl, pH 7.5, 50 mM NaCl with Roche EDTA-free Complete protease inhibitor cocktail). For pull-down experiments with purified protein, 0.3114 nmol of purified bait protein in 100 μL NP-40 buffer was bound to either 50 μL anti-HA agarose (Sigma) or 40 μL anti-FLAG M2 affinity gel (Sigma) and rotated for 5-6 hours 4˚C. The bait-bound matrix was then washed 3 times with NP-40 buffer and twice molar excess (0.6228 nmol) of prey protein was added (unless otherwise indicated) in 275 μL NP-40 buffer and rotated at 4˚C overnight. The matrix was then washed three times and immune complexes were eluted with either 500 μg/ml HA peptide or 150 μg/ml FLAG peptide in IP elution buffer as described above.

Antibodies and western blot
For western blot analyses, FLAG-tagged proteins were detected with mouse monoclonal anti-FLAG clone M2 antibody (Sigma) and HA-tagged proteins were detected with mouse monoclonal anti-HA clone HA-7 (Sigma). To detect DciA Mtb , we used a rabbit polyclonal anti-DciA Mtb antibody that was generated by Cocalico Biologicals, Inc. by raising rabbit anti-sera against purified DciA Mtb protein. Western blots were visualized on a ChemiDoc Touch Imaging System (Biorad) as well as film and quantified using ImageLab software version 5.2.1 software (Biorad).
Chromatin immunoprecipitation quantitative PCR (ChIP-qPCR) 50 mL cultures of M. smegmatis expressing either HA-tagged DciA Mtb (HA-DciA Mtb ), HAtagged CarD (HA-CarD), or untagged DciA Mtb (No tag) were grown to an ODλ 600 of 0.6. Protein-nucleic acid complexes were crosslinked with 2% formaldehyde for 30 min at room temperature, crosslinking was quenched with 0.125M glycine, and cells were lysed in ChIP lysis buffer (50mM HEPES-KOH pH7.5, 140mM NaCl, 1mM EDTA, and 1% TritonX 100, 1X protease inhibitors (Roche), and 2mg/ml lysozyme). After sonication, 0.5 ml of lysate was kept as input fractions and the remaining lysate was rotated with anti-HA agarose (Sigma) overnight at 4˚C. Anti-HA agarose was washed 2 times each with ChIP lysis buffer, ChIP lysis buffer plus 360mM additional NaCl, ChIP wash buffer (10mM Tris-HCl pH8, 250mM LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, and 1mM EDTA), and TE buffer (10mM Tris-HCl pH8 and 1mM EDTA). All buffers were supplemented with 1X protease inhibitors. Proteinnucleic acid complexes were eluted from the anti-HA agarose twice with ChIP elution buffer (50mMTris-HCl pH8, 10mM EDTA, 1% sodium dodecyl sulfate, 1X protease inhibitors) for 10 min at 65˚C with agitation. Inputs and eluates were incubated at 65˚C overnight to reverse crosslinking. Two phenol-chloroform extractions were performed consecutively on inputs and eluates and DNA was precipitated with ethanol and resuspended in TE buffer. Inputs and eluates were diluted to 0.5896 ng/μl and 0.426 ng/μl, respectively. Quantitative PCR (qPCR) was performed with 1 μl of the diluted DNA using primers to amplify three DNA fragments in oriC (oriC1, oriC2, oriC3), a fragment in the promoter of rplN, and a fragment within the sigA coding region (sigAIN) (See S3 Table and

Biolayer interferometry
Recombinant DnaB and DciA Mtb were purified from E. coli as described above. An additional dialysis step to remove Tris from the buffer was performed into 150mM NaCl, 10mM NaPO 4 pH8, 1mM BME. DnaB was non-specifically biotinylated in vitro using EZ-Link NHS-PEG 4 -Biotin (ThermoFisher). Five molar excess of biotin relative to DnaB was used and incubated for 30 minutes at room temperature. Excess biotin was removed using Zeba Spin Desalting columns (ThermoFisher).
The Octet RED96 System was used to attain biolayer interferometry (BLI) progress curves. Assay buffer for all steps consisted of 150mM NaCl, 0.02% Tween-20, 0.1% bovine serum albumin, and 10mM NaPO 4 pH8. Briefly, streptavidin (SA) biosensor pins (ForteBio) were first equilibrated by being dipped into assay buffer for a 180 second baseline step, and then captured 200nM biotinylated DnaB during a 200 second loading step, followed by a 180 second baseline step in assay buffer. After performing an additional 60 second baseline step in assay buffer, pins were dipped into DciA Mtb protein samples for a 300 second association step, followed by a 300 second dissociation step in assay buffer. This series of 60 second baseline, 300 second association, and 300 second dissociation steps was performed for each concentration of DciA Mtb . Curves were corrected by subtracting double reference of both biotin-coated pins dipped into the DciA Mtb wells and DnaB-coated pins dipped into buffer only. Data was analyzed on ForteBio Data Analysis 6.4. Processed data were fit globally for all concentrations of DciA Mtb in a 1:1 kinetic binding model.

Mass spectrometry
Bands generated by SDS-PAGE followed by staining with ProteoSilver Plus Silver Stain Kit (Sigma), were cut out and destained according to manufacturer's instructions, and submitted to the Proteomics & Mass Spectrometry Facility at the Danforth Plant Science Center for trypsin digestion followed by LC-MS/MS analysis.

Statistics
Prism6 (Graphpad Software, Inc.) was used to determine statistical significance of differences. Unpaired two-tailed Student's t-test was used to compare two groups with similar variances. Unpaired two-tailed Student's t-test with Welch's correction was used to compare two groups with different variances. One-way analysis of variance (ANOVA) and Tukey's multiple comparison test were used to determine significance when more than two groups were compared. When utilized, center values and error bars represent mean ± SEM. Ã p <0.05, ÃÃ p <0.01, ÃÃÃ p<0.001, ÃÃÃÃ p <0.0001.