**3. Discussion**

*Brucella abortus* is able to control its cell cycle progression when it is inside host cells, particularly the replication and segregation of its replication origins [2]. However, the molecular mechanisms involved in this control are unknown. Since convincing data show that the ability to adapt to starvation is a key factor for the success of cellular infections by *Brucella melitensis* and *Brucella suis* [10,11], we further investigate the role of the (p)ppGpp, the alarmone produced in the presence of starvation conditions, and the Rsh enzyme that is proposed to synthesize this alarmone. We first confirmed that in *B. abortus* 544, like in other *Brucella* strains, Rsh is crucial for the success of a cellular infection (Figure 3), for the survival in stationary phase (Figure 2) and the growth in minimal medium (Figure 1). In agreemen<sup>t</sup> with the absence of proliferation of the *rsh* mutant in macrophages, a strain constitutively producing a (p)ppGpp hydrolase (Mesh1b) from *Drosophila melanogaster* is also unable to grow in RAW 264.7 macrophages. We observed that the survival of *pBBR-mesh1b* strain is less severely impacted during infection than the Δ*rsh* strain. One could imagine that this intermediate phenotype is due to the presence of residual alarmone in the *pBBR-mesh1b* strain, while it is not the case in the Δ*rsh* strain since the (p)ppGpp synthetase domain is not present. Another explanation for these di fferent phenotypes could be that Rsh plays additional role(s) for survival in infection than the regulation of (p)ppGpp homeostasis. Interestingly, a mutant overproducing (p)ppGpp is also unable to proliferate in these cells, suggesting that the (p)ppGpp level should be in a specific range of concentration to allow cellular infection; having too much or not enough (p)ppGpp would be detrimental for the success of the cellular infection.

What does (p)ppGpp control and how is Rsh regulated? It was shown that the absence of *rsh* in *B. suis* affects the transcription of genes known to be involved in virulence [28], such as *pyrB*, which was shown to be essential for *B. abortus* proliferation in RAW 264.7 macrophages [25]. It is thus likely that a minimum level of (p)ppGpp would be required for the success of a cellular infection by *B. abortus*. It was shown that the glutamine pool modulates Rsh (called SpoT) in the model alpha-proteobacterium *Caulobacter crescentus* through the phosphotransferase system (PTS) [14]. Interestingly, mutants for components of this system were found to be attenuated in RAW 264.7 macrophages [25]. Moreover, PTS and the two-component regulator BvrR control the expression of the *virB* operon [29,30], coding for a type IV secretion system that is crucial for intracellular proliferation in most cell types [31]. These data indicate that a quite complex regulation network is probably linking Rsh control and virulence. However, the molecular mechanisms controlling Rsh activity in *B. abortus* are unknown and deserve further investigation.

One striking conclusion of our data is the moderate e ffect that (p)ppGpp overproduction has on the proportion of bacteria at the G1 stage of the cell cycle. Indeed, while overproduction seems to be su fficient to impair growth inside host cells (Figure 7), the proportion of G1 after 6 h of induction is about a third of the culture while it is approximately 15% in the absence of induction (Figure 6). Since the cell cycle takes about 3 h in the conditions tested, it is likely that only a fraction of the bacteria arrested their cell cycle at the G1 stage. Inside host cells, the proportion of G1 cells is about 75% and remains stable for 2 to 4 h at least [2], suggesting that other mechanisms are probably involved in the control of the cell cycle in host cells, early in the tra fficking. These mechanisms could involve the acidic nature of the BCV, or the di ffusion sensing proposed to occur through a regulation system homologous to quorum sensing [32]. More investigations are thus needed to discover the multiple factors involved in the cell cycle control of *B. abortus* inside host cells.

### **4. Materials and Methods**

### *4.1. Strains and Growth Conditions*

The reference strain *B. abortus* 544 was used for all experiments and was grown on solid or in liquid 2YT medium (LB 32 g/<sup>L</sup> Invitrogen, Yeast Extract 5g/L, BD and Peptone 6 g/L, BD) at 37 ◦C. *E. coli* strain DH10B was used for plasmid constructs and the conjugative strain *E. coli* S17-1 was used for mating with *B. abortus*. Both strains were cultivated in LB medium (Luria Bertani, Casein Hydrolysate 10g/L, NaCl 5g/L, Yeast Extract 5g/L) at 37 ◦C. Depending on the plasmid used, di fferent selection antibiotics were added to the culture medium: ampicillin (100 μg/mL); carbenicillin (100 μg/mL); kanamycin (50 μg/mL for the replicative plasmid, and 10 μg/mL for the integrated plasmid); nalidixic acid (25 μL/mL); chloramphenicol (20 μg/mL for the replicative plasmid and 4 μg/mL for the integrated

plasmid). Isopropyl β-d-1-thiogalactopyranoside (IPTG) was used at a concentration of 1 mM in bacterial liquid culture and at 10 mM in the mammalian cell culture medium during cellular infections. When the Δ*rsh* mutant was constructed, we added casamino acids 0.5% (BactoTM Casamino Acids from Thermo Fisher, Waltham, MA, USA) in the conjugation medium.

### *4.2. Strains Construction*

Deletion strains were constructed by allelic exchange using the pNPTS138 vectors (M. R. K. Alley, Imperial College of Science, London, UK) carrying the upstream and the downstream regions of the targeted gene. The primer sequences used for amplification of the upstream region of the *dksA* gene were 5'-ttGGATCCcaagcgccagatcttca-3' and 5'-ttGAATTCttcactcattctgaatcacccc-3'. The primer sequences used for amplification of the downstream region of the *dksA* gene were 5'-ttGAATTCtgatatcgaataatggtttggaaa-3' and 5'-ttAAGCTTcgcccagcttcaaattac-3'.

We used the *rsh* deletion plasmid pMQ203 (provided by M. Quebatte, Biozentrum, Basel), containing the upstream and downstream regions of *rsh* amplified with the following hybridization sequences: 5'-ccggatgatctgaaggaa-3', 5'-gcgcatcatctgccgaaa-3' and 5'-gtctgggacctcaagcat-3', 5'-cccgtggtgacgatatct-3'.

The Δ*rsh* pBBR-*rsh* strain was generated by inserting the pBBR-*rsh* in the Δ*rsh* strain. The pBBR-*rsh* was constructed by cloning the endogenous promoter of *rsh* and the *rsh* coding sequence, amplified with the primers 5'-aaaCTCGAGgcgagattgccgatgaga-3' and 5'-aaaCTGCAGctatccgttcacacgctttg-3'.

The pBBRi-*mesh1b* strain was constructed by inserting the coding sequence *mesh1b* in the pBBRi plasmid. The sequence of *mesh1b* was adapted to the codon usage of *Brucella*, and is available in Supplementary Figure S2.

The *mCherry-parB* strains containing pSRK-*relA'* and pSRK-*relA'*\* were created using the Tn7 system [33] which consists in transposition of mini-Tn7 expressing *mCherry-parB* under the control of the *PgidA* promoter as previously reported [2] and the resistance cassette to ampicillin/carbenicillin under the control of P*bla* promoter at the *glmS* locus of *B. abortus*. The primer sequences used for the amplification of <sup>P</sup>*gidA*-mCherry-*par<sup>B</sup>* and P*bla-amp* were 5'-cgcggatcctctgtggaatcctgtttgttg-3', 5'-AGCGGATACATATTTGAActagctttgaagacggcg-3' and 5'-TTCAAATATGTATCCGCTCATGA-3', 5'-cgggatccTTACCAATGCTTAATCAGTGAGG-3'.

### *4.3. Growth Assays*

The bacterial growth curves were performed using a bioscreen (Epoch2 Microplate *Photospectrometer* from BioTek). Bacterial cultures in the exponential phase of growth were washed two times with PBS and were normalized at an OD of 0.1 in a given medium. A 200 μL aliquot of the normalized culture was transferred to a plate and each condition was performed in technical triplicate (3× 200 μL). The plate was incubated at 37 ◦C with shaking and the OD600 of each well was measured every 30 min. One biological replicate constitutes the mean of three technical replicates and experiments were repeated at least three times to obtain biological triplicates.

### *4.4. Survival Assays*

Bacterial cultures in the exponential phase of growth were normalized at an OD of 0.1 in 2YT liquid medium. Serial dilutions were plated on 2YT solid medium at different time points and plates were incubated at 37 ◦C.

### *4.5. Infections of RAW 264.7 Macrophages*

RAW macrophages were put in wells in DMEM medium (with decomplemented bovine serum, glucose, glutamine, and no pyruvate, Gibco®) to have 1 × 10<sup>5</sup> cells/mL. *B. abortus* 544 was grown in 2YT at 37 ◦C until exponential phase. The OD of the bacterial culture was measured, and dilutions were performed to have a MOI equal to 50 (50 times more bacteria than macrophages). An input control was performed for each condition by plating bacteria on a 2YT agar plate before infecting cells. Cell medium was removed to add the appropriate bacterial dilution. The mix was centrifuged for 10 min at 1200 rpm (4 ◦C) and then incubated at 37 ◦C with 5% CO2 (this time point is set as time zero). After one hour of incubation, medium was removed and replaced by medium containing gentamycin (50 μg/mL) for 1 h in order to kill extracellular bacteria, and then by medium containing gentamycin (10 μg/mL). Note that for the experiments using IPTG, the IPTG (10 mM) was kept during all the steps of the infection. At either 2 h, 4 h or 24 h post infection, cells were first washed with sterile PBS and were then incubated in PBS + Triton 0.1% at 37 ◦C for 10 min in order to lyse the cells while keeping bacteria alive. After that, cells were flushed and lysates were harvested. Serial dilutions were performed and each dilution was spotted on 2YT agar plates and incubated at 37 ◦C.

### *4.6. Infections of HeLa Cells*

HeLa cells were plated in wells in DMEM medium (with sodium pyruvate, non-essential amino acid, glucose, glutamine, and no pyruvate, Gibco®) at 4 × 10<sup>4</sup> cells/mL. *B. abortus* 544 was grown in 2YT at 37 ◦C until exponential phase, the OD of the bacterial culture was measured, and dilutions were performed to have a MOI equal to 300. An input control was performed for each condition by plating bacteria on a 2YT agar plate before infecting cells. Prior to infections in the presence of IPTG (see below), *relA'* expression was induced 3 h before infection with IPTG (1 mM in YT medium). Cell medium was removed to add the appropriate bacterial dilution. The mix was centrifuged for 10 min at 1200 rpm (4 ◦C) and incubated at 37 ◦C with 5% CO2 (this time point is set as time zero). After one hour of incubation, medium was removed and replaced by medium containing gentamycin (50 μg/mL) in order to kill extracellular bacteria, and then gentamycin (10 μg/mL). Note that for the experiments using IPTG, the IPTG (10 mM) was kept during all the steps of the infection. At either 2 h, 4 h or 24 h post infection, cells were first washed with sterile PBS and were then incubated in PBS + Triton 0.1% at 37 ◦C for 10 min in order to lyse the cells while keeping bacteria alive. After that, cells were flushed and lysates were harvested. Serial dilutions were performed and each dilution was spotted on 2YT agar plates and incubated at 37 ◦C.

### *4.7. G1 Counting*

Bacteria in exponential phase of growth were diluted to an OD of 0.1 in 2YT liquid medium with or without IPTG. At each time point, 200 μL of the culture was washed two times in PBS and bacteria were loaded onto a PBS agarose pad to be observed and counted by fluorescence microscopy.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-0817/9/7/571/s1, Figure S1: Effect of dksA deletion on macrophage infection; Figure S2: Codon-adapted mesh1b sequence; Figure S3: Fluorescence microscopy of the pSRK-relA' mCherry-parB strain induced with IPTG; Table S1: Number of bacteria counted in Figure 5.

**Author Contributions:** M.V.d.H. and X.D.B. designed the work and wrote the manuscript; M.V.d.H. and E.C. performed the experiments. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by FRS-FNRS (PDR T.0060.15 and T.0058.20) and the University of Namur.

**Acknowledgments:** We thank Maxime Quebatte (Biozentrum, Basel) for the generous gift of the pMQ203 plasmid. We thank Severin Ronneau for the generous gift of the pSRK-*relA'* plasmid and for the numerous fruitful discussions on (p)ppGpp. M.V.d.H. was supported by a FRIA Ph.D. fellowship from FRS-FNRS. We thank the University of Namur for logistic support.

**Conflicts of Interest:** The authors declare no competing interests.
