Distinct architectural requirements for the parS centromeric sequence of the pSM19035 plasmid partition machinery

To define the requirements for stimulation of the ParA_pSM ATPase activity by ParB_pSM, the steady-state ParA_pSM ATP turnover rate was measured with varying concentrations of ParB_pSM in the presence or absence of different duplex DNA cofactors. The turnover rate of ParA_pSM ATPase alone is very low at 37°C (0.9 ± 0.1 ATP/ParA-dimer/h, N = 3; Figure 1A; no DNA). In the presence of a saturating concentration of double-stranded DNA (40 μg/ml pBR322 plasmid DNA plus 23–38 μg/ml double-stranded oligonucleotide with or without parS_pSM sequence), to which ATP-ParA_pSM dimers can bind to support ATPase activation by ParB_pSM (see below), no significant rate change was observed (1.0 ± 0.1 h^–1, N = 46, Figure 1A, pool of all measurements at [ParB_pSM] = 0). Next, effects of the parS_pSM-DNA in addition to 40 μg/ml nsDNA and ParB_pSM were examined. ParA_pSM ATP hydrolysis was stimulated up to ~20-fold (k_cat = 20.5 ± 2.9 h^–1, N = 7) in the presence of ParB_pSM and oligonucleotide duplex DNA containing one of the native arrangements containing seven parS_pSM heptad-sequence-repeats (7R-parS_pSM, →→←→→←←) (Figure 1A). The stimulation approached saturation around 2 μM ParB_pSM at varying ParA_pSM concentrations (Figure 1—figure supplement 1A). In contrast, when the parS_pSM DNA fragment was replaced with one having a scrambled sequence, ParB_pSM stimulated ParA_pSM ATPase activity only to 2.4 ± 1.1 h^–1 even in the presence of 8 μM ParB_pSM that should have allowed non-parS_pSM DNA binding (N = 3*) (Figure 1A, scram). Similarly, a low level of ParA_pSM ATPase stimulation was observed with ParB_pSM^1-27 peptide, which lacked the DNA-binding and dimerization domains (Figure 1—figure supplement 1B). These results demonstrated that full ParA_pSM-ATPase stimulation by ParB_pSM requires specific parS_pSM interactions. Even the low parS_pSM-independent ATPase stimulation was not detected in the absence of DNA (Figure 1A, no DNA). Similarly, the low-level ATPase stimulation by non-parS_pSM-binding ParB_pSM^1-27 peptide was not detected without DNA (Figure 1—figure supplement 1B). We conclude ParA_pSM-nsDNA binding is required for the ATPase activation by ParB_pSM, as reported for ParA_F ATPase activation by ParB_F (Taylor et al., 2021). Consistently, in the absence of nsDNA, excess ParB_pSM relative to the concentration of 7R-parS_pSM in the reaction inhibited ATPase stimulation, presumably competing with ParA_pSM for parS_pSM DNA binding (Figure 1—figure supplement 1C).

Figure 1 with 3 supplements see all

Download asset Open asset

Figure 1—source data 1

https://cdn.elifesciences.org/articles/79480/elife-79480-fig1-data1-v2.xlsx

Download elife-79480-fig1-data1-v2.xlsx

Formation of short clusters of ParA_pSM molecules bound to nsDNA might be taken as a hint for cytoskeletal filament treadmilling models for this class of partition systems, rather than diffusion-ratchet model. However, the low maximum ATP turnover rate (~0.3 /min) observed would be too slow for a treadmilling filament model. Combined with low abundance of ParA_pSM molecules inside a cell insufficient to form axial protein filaments, the treadmilling model is highly unlikely to fit this system.

We asked whether entire 7R-parS_pSM is required for the full stimulation of ParA_pSM ATPase by ParB_pSM. A series of deletions of parS_pSM heptad-sequence-repeats were made and their ATPase stimulation activities were tested (Figure 1E). Efficient ATPase stimulation was observed when 6R (→←→→←←), 5R (→→→←←), or 4R (→→←←) was added along ParB_pSM (Figure 1A). No significant difference was observed among 5R, 6R, or 7R; both half-saturation concentrations and the apparent k_cat were comparable (Figure 1A). ParB_pSM-4R-parS_pSM also induced similar rates of ParA_pSM ATP hydrolysis, but a significantly higher concentration was required for full stimulation (Figure 1A). Different heptad orientation arrangements of 4R-parS_pSM (→→←→ and →→→→) stimulated ParA_pSM ATPase to a similar extent (Figure 1B). In contrast, 3R- (→→←), 2R- (→←), and 1R- (→) parS_pSM were poor cofactors for ParB_pSM-dependent ATPase stimulation (apparent k_cat = 3.2–3.6 ± 1.3–2.1 h^–1, N = 4–7*, Figure 1A), not significantly different from the scrambled sequence DNA. 3R-parS_pSM fragments with different repeat arrangements behaved similarly to each other (Figure 1C). The ATPase stimulation by parS_pSM DNA with 1–3 parS_pSM repeats also appeared to saturate at around 2 μM ParB_pSM, indicating that the affinity of the nsDNA-bound ParA_pSM to ParB_pSM in the presence of truncated parS_pSM was not limiting above ~2 μM ParB_pSM. Previously it has been shown that ParB_pSM bound 3R, 4R DNA fragments of different sequence orientation combinations, or a 10R DNA fragment with roughly similar affinity that was >50-fold higher than nsDNA or 1R DNA (de la Hoz et al., 2004). We confirmed that ParB_pSM binding was similarly strong for 4R and 3R duplex DNA (K_D ~ 17 nM) and weak for nsDNA (K_D ~ 1 μM, Figure 1—figure supplement 3). Therefore, the affinity of ParB_pSM for the 3R-parS_pSM DNA is not limiting the ParA_pSM ATPase stimulation. In the above experiment, the length of parS_pSM DNA fragments decreased as the number of parS_pSM repeats decreased (Figure 1E, Figure 1—figure supplement 2). We tested longer 3R-parS_pSM DNA fragments containing an additional non-parS_pSM heptamer sequence (non-consensus, nc) (3R-1nc) and confirmed that parS_pSM-DNA fragment size was not a significant factor for the ATPase stimulation efficiency (Figure 1C).

These results suggested that ParA_pSM ATP turnover is fine-tuned by the structural arrangement of the ParB_pSM and parS_pSM within the PC. ParB_pSM assembles as a left-handed spiral to wrap parS_pSM DNA without significantly distorting the DNA backbone geometry (Weihofen et al., 2006), implying that after approximately four repeats, ParB_pSM dimers would make a full turn around parS_pSM, positioning themselves on the same face of the nucleoprotein filament. Thus, the position of the fourth repeats relative to that of the first repeat might be functionally important. To test this possibility, we examined the ParA_pSM ATPase stimulation efficiency of parS_pSM DNA fragments containing two copies of 3R sequences separated by 7 bp or 14 bp of non-consensus sequence (3R-1nc-3R and 3R-2nc-3R) (Figure 1E). The 3R-1nc-3R-parS_pSM fragments showed only slightly higher stimulation (k_cat = 8.3 ± 3.5 h^–1, N = 8*) compared to the single 3R-parS_pSM fragment, and the 3R-2nc-3R was functionally indistinguishable from the single 3R-parS_pSM fragment (Figure 1C). We conclude that disrupted 6R-parS_pSM fragments are unable to recover the full stimulation of ParA_pSM ATPase activity. These findings indicate that the ParB_pSM-ParA_pSM interactions differ when ParB_pSM dimers are bound to ≥4 contiguous heptad repeats compared to ParB_pSM dimers unbound to the repeats or bound to fewer or a disrupted array of heptad repeats.

We originally assumed that for efficient ParA_pSM ATPase activation by the ParB_pSM-parS_pSM complex all ParB_pSM dimers need to be bound to parS_pSM and accordingly maintained stoichiometrically excess parS_pSM-sequence concentration relative to ParB_pSM-dimers in the reaction. To test this assumption, we next changed the concentration of the 7R-parS_pSM DNA fragment while keeping the concentrations of ParA_pSM and ParB_pSM both at 2 μM. Contrary to our expectation, significantly lower heptad consensus sequence concentrations of the 7R-parS_pSM DNA fragment compared to ParB_pSM dimers (B₂) were sufficient for ATPase activation (Figure 1D). If the original assumption was correct, 2 μM ParB_pSM should have required 1 μM consensus sequence repeats (143 nM 7R-parS_pSM) for full activation. According to the results of Figure 1A, at high 7R-parS_pSM concentration, half-activation by ParB_pSM required ~650 nM ParB_pSM, which would have required ~325 nM parS_pSM consensus sequence if full binding was needed. Observed half-saturation parS_pSM repeat concentration was ~130 nM (~19 nM 7R-parS_pSM) or less. Thus, assuming most of the ParB_pSM molecules in our preparations are active, it appears that at any given time, less than half of the ParB_pSM dimers need to be in complex with 7R-parS_pSM to exert full ATPase activation.

Next, we examined how the dissociation of ParA_pSM-ATP dimers from nsDNA is influenced by ParB_pSM in the presence of parS_pSM DNA. For these experiments, we used an nsDNA-carpeted two-inlet flow cell observed under TIRF microscopy as described in ‘Materials and methods’ (Vecchiarelli et al., 2013; Figure 2A). We used a ParA_pSM-GFP fusion protein and a ParB_pSM-cys conjugated to Alexa647 to facilitate visualization of protein association-dissociation to/from the nsDNA-carpet. Stimulation of ParA_pSM-GFP ATPase activity by ParB_pSM and 7R-parS was comparable to wild-type ParA_pSM (Figure 2—figure supplement 1). First, we tested the association/dissociation of individual proteins (bound at 1 μM) to/from the nsDNA-carpet. The affinity of ParB_pSM-Alexa647 (1% labeled) to nsDNA was weak. ParB_pSM-Alexa647 in the absence of ParA_pSM accumulated on the DNA-carpet reaching a density of ~3000 ± 500 dimers/μm² after 15 min of constant flow (5 μl/min) (Figure 2—figure supplement 2A, bottom). When the sample solution was switched to a buffer without protein, most ParB_pSM rapidly dissociated with kinetics that can be fitted to a double-exponential function (k_off-fast = 5.1 ± 0.7 min^–1 [73%], k_off-slow = 0.11 ± 0.04 min^–1, N = 3; all k_offs reported in this study represent apparent pseudo-first order rate constants) (Figure 2—figure supplement 2B, bottom). ParA_pSM-GFP preincubated with ATP and Mg²⁺ reached a saturation density on the nsDNA-carpet of 4.25 (±0.06) × 10⁴ dimers/μm² (N = 4*), whereas in the absence of ATP, less than 1000 dimers/μm² of ParA_pSM bound to the nsDNA-carpet (Figure 2—figure supplement 2A, top). Next, 1 μM ParA_pSM-GFP preincubated with ATP was flowed onto the nsDNA-carpet until 5–10% of saturation density (~4000 dimers/μm²), at which point the flow was switched to buffer containing ATP without ParA_pSM-GFP. A small fraction of ParA_pSM-GFP (less than ~5%) dissociated within ~1 min and the remainder dissociated slowly with k_off of <0.013 min^–1 (N = 3, Figure 2B).

Figure 2 with 3 supplements see all

Download asset Open asset

Figure 2—source data 1

https://cdn.elifesciences.org/articles/79480/elife-79480-fig2-data1-v2.xlsx

Download elife-79480-fig2-data1-v2.xlsx

The addition of competitor nsDNA (71 nM scrambled 55 bp duplex DNA) in the wash buffer slightly increased the k_off of majority (~93%) fraction of ParA_pSM-GFP (k_off = 0.075 ± 0.001 min^–1, N = 3), perhaps in part by competing with rebinding of ParA_pSM-GFP-ATP to the nsDNA-carpet (Figure 2B). When 1 μM ParB_pSM (unlabeled or 1% Alexa647 labeled) was added to the wash buffer without competitor nsDNA, k_off of ParA_pSM-GFP from nsDNA-carpet reached 0.17 ± 0.01 min^–1 (N = 4, Figure 2B). ParA_pSM-GFP dissociation from the nsDNA-carpet was accelerated when wash solution contained 71 nM 7R-parS_pSM DNA (0.5 μM of the consensus sequence repeat) and 1 μM ParB_pSM (k_off = 1.97 ± 0.09) after a brief period of initial slower dissociation rate (see below for details) (Figure 2B). This was ~12 or ~26-fold higher k_off compared to ParB_pSM or nsDNA wash, respectively. Similarly, 4R-parS_pSM DNA supported accelerated ParA_pSM-GFP release by ParB_pSM, but 3R-parS_pSM DNA did not (Figure 2B). The results paralleled those shown in Figure 1A solidifying the notion that a functional parS_pSM must carry four copies of ParB_pSM dimer-binding sequence repeats.

The dependence of ParA_pSM dissociation behaviors on the heptad repeat number of the parS_pSM fragment combined with 1 μM ParB_pSM in the wash solution when the wash was started with ParA_pSM bound to the nsDNA-carpet at a saturating density was qualitatively similar. However, the dissociation kinetics were significantly slower (Figure 2—figure supplement 3A), likely reflecting the approximately tenfold higher starting ParA_pSM-GFP density on the nsDNA-carpet, for which the concentration of ParB_pSM used likely was limiting.

The accelerated ParA_pSM-GFP release from the nsDNA-carpet when washed with ParB_pSM-4R or –7R parS_pSM complexes started after an initial slower rate of ParA_pSM release (Figure 2B). This time lag prompted us to examine the binding kinetics of ParB_pSM to the ParA_pSM-bound nsDNA-carpet at different concentrations of ParB_pSM. Upon switching the flow to the washing solution without parS_pSM, ParB_pSM bound to the ParA_pSM-bound nsDNA-carpet to a peak density of ~3000 dimers/μm² and then quickly decreased to a plateau density of ~2000 dimers/μm² (Figure 3Ad). The initial overshoot of ParB_pSM binding appeared to roughly coincide with the fast dissociation phase of ParA_pSM. During the subsequent slow dissociation of ParA_pSM, ParB_pSM density on the nsDNA-carpet remained relatively constant with the protein ratio reaching ~1 after several min of washing with 1 μM ParB_pSM (Figure 3Ad).

Figure 3

Download asset Open asset

Figure 3—source data 1

https://cdn.elifesciences.org/articles/79480/elife-79480-fig3-data1-v2.xlsx

Download elife-79480-fig3-data1-v2.xlsx

In the presence of 1 μM ParB_pSM and 7R-parS_pSM in the wash solution, ParA_pSM-GFP release from the nsDNA-carpet took place in two phases (Figure 3Aa). First, ParB_pSM accumulated on the ParA_pSM-bound nsDNA-carpet to a density twofold or more in excess of the carpet-bound ParA_pSM. During this phase, ParA_pSM dissociated from the nsDNA-carpet relatively slowly. Next, as the binding of ParB_pSM stopped and began to quickly dissociate (k_off = 2.4 ± 0.1 min^–1, N = 5), ParA_pSM dissociation also accelerated to a higher rate (k_off = 1.8 ± 0.05 min^–1, N = 5) (Figure 3Aa). Near-complete dissociation of both proteins from the nsDNA-carpet occurred within a few minutes. Biphasic dissociation kinetics of ParA_pSM was clearer when the wash solution contained lower concentrations of the 7R-parS_pSM-ParB_pSM complex, as ParB_pSM accumulated on the ParA_pSM-bound nsDNA-carpet slower, and it took longer to transition to the accelerated dissociation phase (Figure 3Ae,i). This indicates that the initial nsDNA-carpet-bound ParA_pSM-ParB_pSM complex is not activated for ParA_pSM dissociation from nsDNA, and slow transition of the nsDNA-bound complex is necessary to start the ParA_pSM disassembly.

We next examined the ability of 4R- and 3R-parS_pSM DNA to accelerate ParA_pSM-GFP dissociation in the presence of ParB_pSM. Switching 7R-parS_pSM in the wash solution to 4R-parS_pSM (1 μM ParB_pSM, 125 nM 4R DNA) yielded qualitatively similar two-step dissociation curves (Figure 3Ab). The k_off of ParA_pSM for the accelerated dissociation phase (1.9 ± 0.06 min^–1, N = 5) was roughly the same as in the presence of 7R-parS_pSM. After the peak of binding, ParB_pSM dissociated from the nsDNA-carpet with k_off of 3.9 ± 0.2 min^–1, N = 5. The dissociation burst of ParA_pSM-GFP was triggered in the presence of 4R-parS_pSM at ParB_pSM/ParA_pSM molar ratio of ~1.7 compared to ~2.8 in the presence of 7R-parS_pSM (at 1 μM ParB_pSM). This difference likely in part reflects the fact that an individual ParB_pSM-parS_pSM complex with 4R-parS_pSM contains fewer ParB_pSM dimers than the 7R-parS_pSM complex contains. Also, at a given ParB_pSM concentration in the wash solution, washing with 7R-parS_pSM took roughly twice as long to reach the peak ParB_pSM/ParA_pSM ratio to start the rapid ParA_pSM dissociation phase compared to the 4R-parS_pSM wash (Figure 3B). We believe this likely reflects, in part, a lower concentration of the larger ParB_pSM-7R-parS_pSM complex than the complex with 4R-parS_pSM. The accelerated phase of ParA_pSM k_off appeared higher in the presence of higher concentrations of 7R- or 4R-parS_pSM-ParB_pSM complex in the wash solution (Table 1).

Unlike the results above using 7R- or 4R-parS_pSM in the wash solution, washing experiments with 3R-parS_pSM (1 μM ParB_pSM, 166 nM 3R-parS_pSM) did not exhibit an accelerated ParA_pSM-GFP dissociation phase, as in the absence of parS_pSM (Figure 3Ac,d). ParB_pSM quickly accumulated onto the ParA_pSM-bound nsDNA-carpet up to a ParB_pSM/ParA_pSM ratio of ~1.3 (1 μM ParB_pSM). Thus, ParB_pSM appears to be able to associate with nsDNA-bound ParA_pSM to a stoichiometry, which is unlikely to be the factor preventing stimulation of ParA_pSM dissociation. Both ParA_pSM-GFP and ParB_pSM dissociated from their peak density on the carpet with double-exponential kinetics roughly in parallel (ParA_pSM-GFP; k_off-fast = 3.8 min^–1 ± 0.4 min^–1 [~19%], k_off-slow = 0.07 ± 0.003 min^–1: ParB_pSM; k_off-fast = 1.4 ± 0.3 min^–1 [~24%], k_off-slow = 0.05 ± 0.004 min^–1, N = 3) (Figure 3Ac).

We next asked whether the accelerated dissociation of ParA_pSM from the nsDNA-carpet in the presence of parS_pSM fragments depends on ATP hydrolysis by ParA_pSM. Among ParA_pSM homologues, a conserved Asp in the Walker A’ motif is critical for ATP hydrolysis (Pratto et al., 2008; Park et al., 2012). We prepared the ParA_pSM^D60E-GFP fusion protein; it exhibited no significant residual ATPase activity and faster ATP-dependent nsDNA binding compared to the wild-type protein (Figure 2—figure supplement 1, Figure 2—figure supplement 2A, top). The apparent k_off of the majority fraction (>80%) of nsDNA-bound ParA_pSM^D60E-GFP washed with ATP-containing buffer (0.015 ± 0.007 min^–1, N = 2) was not significantly different from ParA_pSM-GFP (0.014 ± 0.002 min^–1, N = 2) (Figure 2—figure supplement 2B, top). However, when the wash solution contained 250 nM ParB_pSM and 36 nM 7R-parS_pSM, ParA_pSM^D60E-GFP dissociation from the nsDNA-carpet was only a factor of ~2 faster (Figure 3Ca; k_off = 0.029 ± 0.005 min^–1, N = 2) than washing with ATP-buffer. As washing started, the ParB_pSM associated with the carpet-bound ParA_pSM^D60E-GFP to a density well beyond that of ParA_pSM^D60E-GFP with similar association kinetics as with the carpet-bound ParA_pSM-GFP (Figure 3Ae). The ParB_pSM/ParA_pSM ratio bound to the nsDNA-carpet at the initial ParB_pSM overshoot peak was ~1.75 in the presence of 7R-parS_pSM, ~1.3 with 4R-parS_pSM, and ~0.6 with 3R-parS_pSM (Figure 3D). After the peak density, ParB_pSM dissociated from the nsDNA-carpet slowly (k_off = 0.036 ± 0.003 min^–1, N = 2) (Figure 3Ca). Thus, the accelerated DNA dissociation of ParA_pSM by the active ParB_pSM-parS_pSM complex is coupled to ATP hydrolysis, and in the absence of hydrolysis, ParA_pSM^D60E-ATP dimers stayed stably on nsDNA while associated with active parS_pSM-ParB_pSM complex.

In the experiments of Figures 2 and 3, ParA_pSM-ATP dimers were bound to the nsDNA-carpet first without ParB_pSM, and during the wash phase of the experiments, the carpet-bound ParA_pSM was constantly exposed to a solution containing ParB_pSM and parS_pSM. Because ParB_pSM and parS_pSM dissociating from carpet-bound ParA_pSM could be exchanged by those in the wash solution, the data did not report the stability of ParA_pSM-ParB_pSM interaction on the nsDNA-carpet. In the next set of experiments, we preincubated an equimolar ratio of ParA_pSM-GFP (1 μM), ParB_pSM (1 μM), and 7R-, 4R-, or 3R-parS_pSM (0.5 μM total concentration of the consensus sequence repeats) in the presence of ATP for 30 min at room temperature. The samples were then infused into the nsDNA-carpeted flow cell at 5 μl/min for 15 min. At this point, the incoming components interacting with the nsDNA-carpet would be approaching a steady state. The carpet-bound protein complexes were then washed with a buffer containing ATP without ParA_pSM, ParB_pSM, or parS_pSM. ParA_pSM-GFP-ATP in the absence of ParB_pSM reached saturation on the nsDNA-carpet in 5–10 min and then was released from the carpet slowly during the wash phase with dissociation kinetics that fit a double exponential function, majority fraction dissociating slowly (k_off-fast=1.7 ± 1.1 min^–1 [~7%], k_off-slow=0.014 ± 0.002 min^–1, N = 2) (Figure 2—figure supplement 2B, top) as observed starting with lower ParA_pSM-GFP density (Figure 2B).

When ParA_pSM-GFP was mixed with ParB_pSM and ATP in the absence of parS_pSM and infused into the flow cell, ParA_pSM-GFP-ATP reached near saturation density on the nsDNA-carpet in ~10 min. During the wash phase, ParA_pSM-GFP dissociated from the nsDNA-carpet with double exponential kinetics (k_off-fast = 2.3 ± 0.5 min^–1, k_off-slow = 0.057 ± 0.01 min^–1, N = 4) as in the absence of ParB_pSM except for the several fold faster slow phase k_off and increased fraction of faster-dissociating population (~30%) (Figure 4Ad). When the reaction mixture also contained 3R-parS_pSM, ParA_pSM-GFP dissociation kinetics were similar to the above, except reduced fraction of the faster dissociating ParA_pSM-GFP population (Figure 4Ac). In these experiments, ParB_pSM first accumulated on the nsDNA-carpet to roughly two-thirds of the density of ParA_pSM. When washing started, the majority (~75%) of ParB_pSM dissociated quickly (~4 or ~7 min^–1, with or without 3R-parS_pSM). Thus, even in the presence of 3R-parS_pSM fragment, most ParB_pSM dissociates from ParA_pSM bound to nsDNA-carpet within half of a minute (Figure 4Ac and B).

Figure 4

Download asset Open asset

Figure 4—source data 1

https://cdn.elifesciences.org/articles/79480/elife-79480-fig4-data1-v2.xlsx

Download elife-79480-fig4-data1-v2.xlsx

When ParA_pSM-GFP and ParB_pSM were preincubated with 7R-parS_pSM, the steady-state ParA_pSM accumulation on the nsDNA-carpet was suppressed by nearly 50%, and ParB_pSM and ParA_pSM accumulated on the nsDNA-carpet at ~1A:1.2B ratio. When the wash was started, the two proteins dissociated from the carpet maintaining roughly constant protein stoichiometry with kinetics that can be fit to a double-exponential curve (k_off-fast-A [~73%] 0.92 ± 0.18 min^–1, k_off-slow-A 0.075 ± 0.04 min^–1, and k_off-fast-B [~70%] 1.1 ± 0.14 min^–1, k_off-slow-B 0.13 ± 0.03 min^–1, N = 4; Figure 4Aa). Thus, before ATP hydrolysis disassembles ParA_pSM dimer from the nsDNA-carpet, ParA_pSM-ParB_pSM protein interaction appears to be significantly more stable in the presence of 7R-parS_pSM compared to the complex involving 3R-parS_pSM (Figure 4B).

In the presence of 4R-parS_pSM, carpet binding and dissociation dynamics of the two proteins were qualitatively similar to those in the presence of 7R-parS_pSM, except for the following differences: 4R-parS_pSM did not suppress ParA_pSM binding to the nsDNA-carpet as effectively as 7R-parS_pSM, the ParB_pSM/ParA_pSM ratio at the end of 15 min sample infusion was ~0.75, which dropped to ~0.5 within 1 min of washing after which the ratio remained constant. Observed k_off were ParA_pSM (k_off-fast-A [60%] 0.67 ± 0.14 min^–1, k_off-slow-A 0.17 ± 0.05 min^–1; and k_off-fast-B [66%] 1.9 ± 0.2 min^–1, k_off-slow-B 0.23 ± 0.02 min^–1, N = 4; Figure 4Ab). It appears that ParA_pSM bound to the nsDNA-carpet interacts slightly less stably with ParB_pSM in the complex assembled with 4R-parS_pSM compared to those assembled with 7R-parS_pSM, but it was still significantly more stable compared to the majority of the ParA_pSM-ParB_pSM complexes that accumulate in the presence of 3R-parS_pSM or without parS_pSM DNA.

In this experiment, the absence of ParB_pSM and 7R- or 4R-parS_pSM in the wash solution flowing over the nsDNA-carpet appeared to have caused a slowdown of dissociation of the ParA_pSM-ParB_pSM complexes from nsDNA compared to the experiments of Figures 2B and 3A. This prompted us to examine whether partial loss of the parS_pSM DNA and/or ParB_pSM from the complex during the washing caused this kinetic change. We repeated the experiments of Figure 4Aa with the addition of ParB_pSM (1 μM) and/or 7R-parS_pSM DNA fragment (500 nM consensus repeats) in the washing solution. The addition of ParB_pSM and 7R-parS_pSM in the washing solution significantly accelerated ParA_pSM k_off to 2.5 ± 0.25 min^–1 (N = 3), a slightly higher level than seen in Figure 3Aa (1.8 ± 0.05 min^–1), which might have been an underestimate due to the preceding slow complex assembly steps. The addition of ParB_pSM or 7R-parS_pSM separately to the washing solution did not have significant effects on the ParA_pSM dissociation kinetics (Figure 4C). Thus, partial loss of ParB_pSM-parS_pSM complex from the nsDNA-bound ParA_pSM at the onset of the wash (7R- and 4R-parS_pSM) and also during the wash (4R-parS_pSM) appears to have compromised the efficiency of ParA_pSM dissociation from nsDNA in the experiments of Figure 4Aa,b. This indicates that not all the ParB_pSM-parS_pSM complexes that can contribute to dissociation of ParA_pSM from nsDNA are stably retained via ParA_pSM to the nsDNA-carpet through the course of reaction.

When preformed ParA_pSM-ParB_pSM complexes bound to the nsDNA-carpet in the presence of 7R- or 4R-parS_pSM DNA were washed with buffer containing ATP-activated ParA_pSM-GFP (0.5 μM), freshly arriving ParA_pSM-ATP dimers on the nsDNA-carpet replacing the dissociating ParA_pSM partially stabilized ParB_pSM (Figure 5a, b). This observation indicates that when ParA_pSM dissociates from nsDNA following ATP hydrolysis, some fraction of ParB_pSM in the presence of 7R- or 4R-parS_pSM that induced ATP hydrolysis can capture the newly arriving ParA_pSM-ATP dimers on the nsDNA-carpet. The complex containing 7R-parS_pSM was more efficient for recapture than the 4R-parS_pSM complex. We propose larger numbers of ParB_pSM dimers complexed with parS_pSM with more sequence repeats can interact simultaneously with multiple ParA_pSM-nsDNA mini-filaments, more readily capturing newly arrived ParA_pSM-ATP dimers at nearby nsDNA sites. Evidence supporting such multivalent interactions between nsDNA-bound ParA_pSM and parS_pSM-bound ParB_pSM has been reported previously (Pratto et al., 2009). ParB_pSM stabilization by replenishment of ParA_pSM-ATP was not observed in the presence of 3R-parS_pSM or in the absence of parS_pSM, consistent with the notion that ParA_pSM-ParB_pSM interactions under these conditions are unstable (Figure 5c,d).

Figure 5

Download asset Open asset

Figure 5—source data 1

https://cdn.elifesciences.org/articles/79480/elife-79480-fig5-data1-v2.xlsx

Download elife-79480-fig5-data1-v2.xlsx

The parS_pSM sequence features required for the function of ParB_pSM, an RHH-ParB protein contrast with those for the HTH-ParB proteins. Whereas natural parS_F sequence for the F-plasmid is composed of 12 repeats of the ParB_F dimer binding sequence, a single copy of this consensus sequence is able to support faithful plasmid partition (Biek and Shi, 1994). Further, HTH-ParB proteins can spread around parS sites, forming large PCs containing many ParB molecules, most of which have moved away from parS sequence. In contrast, RHH-ParB proteins lack the CTPase domain that is critical for HTH-ParB spreading (Soh et al., 2019; Jalal et al., 2020; Osorio-Valeriano et al., 2019) and they cannot spread from parS sites to flanking DNA (Pratto et al., 2009). Systems involving RHH-ParB proteins perhaps evolved a fundamentally different system architecture for PC dynamics to accomplish robust partitioning of replicated plasmid copies. It is currently unclear what constitutes the defining feature of the ParB_pSM molecules bound to four or more contiguous parS_pSM heptad-sequence-repeats: we offer a possible scenario below. It is interesting to note that parS_pSM-dependence of ParA_pSM-ATPase activation is high compared to F-plasmid Par system involving HTH-ParB, which exhibits either only less than twofold higher ATPase activation in the presence of parS_F (in the absence of CTP) or no difference of the maximum ATP turnover rate in the presence of CTP (Ah-Seng et al., 2009; Taylor et al., 2021). This is reasonable for HTH-ParB systems; most ParB molecules in the PC are not parS-associated at the time it interacts and activates the ParA_F-ATPase and only limited number of ParB_F molecules are outside of PCs. In contrast, fewer PC-associated RHH-ParB molecules and perhaps comparatively more non-PC-associated RHH-ParB molecules exist inside a cell, necessitating tighter parS-dependent ATPase activation for the nonspreading ParB systems.

ParA_pSM-ATP is slow to dissociate from the nsDNA-carpet in the presence of ParB_pSM alone, or ParB_pSM-3R-parS_pSM complex in the wash solution. We showed that ParB_pSM can quickly interact with nsDNA-bound ParA_pSM in the absence of functional parS_pSM DNA (Figure 3Ac). However, this ParA_pSM-ParB_pSM interaction does not fully activate the ATPase (Figure 1). All wash curves of ParA_pSM-ATP dimers bound to nsDNA-carpet under conditions that do not trigger efficient ATPase activation exhibited double-exponential dissociation kinetics, although the small amplitude of the fast dissociation phase, typically ~5%, made quantitative comparison difficult (Figures 2B and 3A). The fast dissociation phase had a time scale of less than 1 min, while the majority fraction dissociated much slower under conditions that do not activate the ATPase. This indicated that the carpet-bound ParA_pSM dimers were composed of two distinct state populations, transitions between which are slow. ParB_pSM with or without 3R-parS_pSM, after the initial rapid binding to nsDNA-bound ParA_pSM, also dissociated with double-exponential kinetics; a fraction of the carpet-associated ParB_pSM quickly dissociated within 1 min before settling to a quasi-steady state with the constant supply of ParB_pSM or ParB_pSM-3R-parS_pSM complex in the wash solution and slowly dissociating ParA_pSM on the nsDNA-carpet (Figure 3Ac,d). We speculate that this small initial binding overshoot reflects ParB_pSM association to the less populated fast-dissociating fraction of the carpet-bound ParA_pSM dimers, which presumably are in a state with faster ParB_pSM-association rate, lower nsDNA affinity, and perhaps closer to the ATP hydrolysis-competent state.

When ParA_pSM-ATP-ParB_pSM complexes bound to the nsDNA-carpet in a steady state with or without 3R-parS_pSM was washed with a simple buffer (Figure 4Ac,d), a double-exponential dissociation of ParA_pSM was again observed with increased fraction (up to ~30%) of the fast-dissociating population. This supports the notion that the fast-dissociating ParA_pSM population is in a state closer to, but not committed to, ATP hydrolysis. We hypothesize that this ParA_pSM transitory state becomes more populated during preincubation with ParB_pSM on the nsDNA-carpet accounting for the moderate stimulation of the DNA-bound ParA_pSM ATPase by ParB_pSM or ParB_pSM-3R-parS_pSM complex at steady state (Figure 1A). However, this ‘fast’ dissociation does not depend on ATP hydrolysis, considering the ATPase-defective mutant ParA_pSM^D60E also exhibits clear double-exponential dissociation (Figure 2—figure supplement 2B, top, Figure 2—figure supplement 3B).

When ParA_pSM-ATP prebound to nsDNA-carpet in the flow cell was washed with a solution containing fully functional 7R-parS_pSM-ParB_pSM complexes, accelerated release of ParA_pSM from nsDNA was observed (Figure 2B, Figure 3Aa,e,i). However, the initial ParA_pSM dissociation kinetics resembled the ParB_pSM wash without fully active parS_pSM and accelerated disassembly started with a delay. This clear transition of the dissociation mode indicates the initial complex between the nsDNA-bound ParA_pSM and parS_pSM-bound ParB_pSM does not immediately trigger ATP hydrolysis. Rather a series of steps must take place subsequent to the initial association of the two protein-DNA complexes to trigger ATP hydrolysis and complex disassembly. This process appears to involve participation of additional ParB_pSM dimers and/or parS_pSM beyond the initial complex, considering that the delay time before the accelerated ParA_pSM dissociation phase depends on their concentration in the wash solution while clear biphasic kinetic feature was observed even at limiting ParB_pSM concentrations.

The ParA_pSM-ParB_pSM interaction is not stable in the absence of 7R- or 4R- parS_pSM (Figure 4B). In contrast, interaction between nsDNA-bound ParA_pSM and ParB_pSM in the presence of 7R-parS_pSM is significantly more stable prior to ATP hydrolysis-dependent dissociation of ParA_pSM from nsDNA (Figure 4Aa). Thus, ParA_pSM-ATP interacts with ParB_pSM in the presence of fully active parS_pSM in a distinct manner than in its absence. For simplicity, we propose this transition to stably interacting ParA_pSM-ParB_pSM complex bound to nsDNA is the limiting step necessary before ATP hydrolysis-dependent acceleration of ParA_pSM dissociation. Slower k_off of ParA_pSM from nsDNA in the experiments of Figure 4Aa,b compared to those in Figure 2B and Figure 3Aa,b, and recovery of faster k_off by the replenishment of ParB_pSM and parS_pSM in the wash solution (Figure 4C) indicated that ParA_pSM-ParB_pSM complex must turn over to maintain high ParA_pSM k_off. This suggests that not all the nsDNA-bound ParA_pSM dimers are interacting with ParB_pSM in the state committed for ATPase activation in the initial set of active complexes that form on the nsDNA-carpet.

Without further experimental constraints, our consideration of the mechanism of ParA_pSM-ATPase activation by ParB_pSM remains speculative. Unlike other members of ParA-family of ATPases studied before, such as ParA_F (Taylor et al., 2021) or MinD (Vecchiarelli et al., 2016), simple interaction of ParB_pSM-N-terminal domain, ParB_pSM^1-27, with ParA_pSM dimers did not efficiently activate the ATPase (Figure 1—figure supplement 1B). Because of the helical nature of the ParB_pSM-parS_pSM complex (Weihofen et al., 2006) and ParA_pSM-nsDNA mini-filament formation (Pratto et al., 2009), two ParA_pSM dimers within a mini-filament might be approximately in position to interact with ParB_pSM dimers separated by two intervening parS_pSM sequence repeats on one face of the helical filament. Let us assume, however, that the pitch and/or the angular arrangements of these two pairs of protein dimers are not in perfect match allowing only partial interaction between ParA_pSM dimers and ParB_pSM dimers in the basal state structures of the two mini-filaments (Figure 6A). Then, the establishment of divalent interactions and full engagement between the two protein-DNA complexes perhaps would force distortions of the structures of both of the protein-DNA mini-filaments (Figure 6B). The force imposed upon ParA_pSM-mini-filament might act as a steppingstone for the conformational change of ParA_pSM necessary for ATPase activation. Such a scenario explains the requirement for the fourth parS_pSM consensus sequence repeat for the ATPase activation. We propose the specific mechanical properties of the contiguous ParB_pSM-parS_pSM mini-filament, likely different from gapped mini-filaments, is required for efficient ATPase activation, explaining the requirement for the parS_pSM sequence repeat contiguity. The divalent interactions between one ParB_pSM-parS_pSM mini-filament with a ParA_pSM-nsDNA mini-filament, with a non-interacting middle segment separating the two interacting pairs, might allow the ParB_pSM-parS_pSM complex to semi-processively activate multiple ParA_pSM-nsDNA complexes in succession by an inchworm-like transfer to a new ParA_pSM-nsDNA mini-filament. The ability of the otherwise dissociating 7R-parS_pSM-ParB_pSM complex after ATPase activation and dissociation of one partner ParA_pSM mini-filament to recapture a freshly arriving ParA_pSM-ATP dimers on the nsDNA-carpet (Figure 5a) is consistent with such a possibility.

Figure 6

Download asset Open asset

The concentration mismatch between ParB_pSM and parS_pSM needed for full activation of ParA_pSM ATPase (Figure 1D) suggests additional details; the functional parS_pSM DNA might be needed only to deliver ParB_pSM dimers onto ParA_pSM-nsDNA mini-filament generating a complex committed to ATP hydrolysis. After ParB_pSM delivery, parS_pSM might release the delivered ParB_pSM dimers, perhaps helped by the mini-filament distortion discussed above, without waiting for ATP hydrolysis (Figure 6C). The emptied consensus sequence on the parS_pSM would then be quickly recharged with free ParB_pSM dimers in solution to activate another ParA_pSM-nsDNA complex, explaining the higher concentration demand for ParB_pSM over parS_pSM for ATPase activation. The requirement for parS_pSM and ParB_pSM replenishment from the washing solution to sustain maximum k_off of the nsDNA-bound ParA_pSM dimers (Figure 4C) supports the notion that the initial active ParA_pSM-ParB_pSM-parS_pSM complex assembled on the nsDNA perhaps does not induce ATP hydrolysis of all the ParA_pSM dimers within the complex. Full disassembly of the complex likely involves release of the parS_pSM fragment, which gets reloaded with ParB_pSM in solution and returns to the ParA_pSM dimers left on the nsDNA. This local recycling of ParB_pSM and parS_pSM would become inefficient when ParB_pSM and parS_pSM are removed from reaction by the buffer wash. We note that dependence of the accelerated phase of ParA_pSM k_off on the concentration of the parS_pSM-ParB_pSM complex (Table 1) is consistent with the above observation. In vivo, local off-rate of ParA_pSM from nucleoid where a PC is located is likely to be significantly faster than observed here, considering the local density of the nucleoid-bound ParA_pSM near a PC would be lower and the ParB_pSM-parS_pSM concentration of a PC, with six copies of parS sites, much higher compared to the conditions in this study.

Earlier we described our model considering one ParB_pSM-parS_pSM mini-filament acting on one ParA_pSM cluster on nsDNA for simplicity. However, this does not immediately explain the clearly biphasic disassembly kinetics of the nsDNA-bound ParA_pSM during washing with ParB_pSM-parS_pSM complex (Figure 2B, Figure 3Aa,e,i). This kinetics is highly reminiscent of activation and membrane dissociation kinetics of membrane-bound MinD-ATPase, a member of ParA-ATPase family, when washed by MinE-containing buffer (Vecchiarelli et al., 2016), which has been proposed to reflect requirement for two MinE dimers for the activation of one MinD dimer. Thus, we suspect two ParB_pSM-parS_pSM mini-filaments may need to cooperate from two sides of one ParA_pSM cluster on nsDNA, and the second binding is kinetically limiting.

Lastly, irrespective of the details of the model, we need to pay attention to a few aspects of the pSM19035 Par system in the context of the diffusion-ratchet concept. As mentioned in the ‘Introduction,’ plasmid or chromosome partition by a diffusion-ratchet mechanism requires the balance between the stability of each ParA/ParB-mediated cargo-nucleoid bridge and the number of these bridges per cargo. In the pSM19035 system, prior to ATPase-activating nsDNA-ParA_pSM-ParB_pSM-parS_pSM complex assembly, the ParA_pSM-ParB_pSM link in the bridge is expected to have a stability close to that in the presence of inactive 3R-parS_pSM (k_off = ~4 min^–1, Figure 4Ac). This stability is comparable to that of F-plasmid ParA_F- ParB_F bridge between the nucleoid and cargo, which we believe to involve one ParA_F dimer, and therefore DNA-carpet-bound ParA_F FRAP rate in the presence of ParB_F (~4 min^–1, Vecchiarelli et al., 2013) would approximate the stability of the bridge. In contrast, we believe that the effective stability of a unit of the cargo-nucleoid bridge by the ATPase-activating nsDNA-ParA_pSM-ParB_pSM-parS_pSM complex containing many ParA_pSM dimers and ParB_pSM dimers is likely to be significantly higher. If one assumes processive rounds of ATPase activation for subsets of ParA_pSM dimers within a complex as considered in our model, individual cargo-nucleoid bridge would have substantially longer lifetime than individual ParA_pSM dimer dissociation rate constant of up to ~2 min^–1 observed by the experiments of Figure 3. At the same time, the number of individual units of bridges to the nucleoid per cargo would be very small compared to up to a few hundred in cases of HTH-ParB plasmid systems. Long-range transfer of the PC from one ParA_pSM cluster to another via simultaneous multiple ParA_pSM cluster interactions involving nucleoid DNA looping without bridge dissolution discussed above would further extend the lifetime of individual bridges, enhancing the effective PC diffusion suppression capability of very small number of bridges without blocking the PC motion. In addition, very slow dissociation of ParA_pSM dimers from nsDNA prior to formation of fully active complex with a PC (k_off < 0.1 min^–1, Figures 2 and 3) compared to ParA_F or ParA_P1 (k_off = ~5 min^–1, Hwang et al., 2013, Vecchiarelli et al., 2013) suggests their very slow diffusion on the nucleoid (practically no hopping). Together with expected slow overall ATP hydrolysis rate, it predicts slow ParA_pSM depletion zone development and refilling. While these parameter shifts compared to the F- or P1-Par systems are generally speaking in the mutually compensatory directions (Hu et al., 2015; Hu et al., 2017), whether they are balanced or not needs to be carefully evaluated as more details of the mechanism come to light in future.

This study revealed a puzzlingly unique parS_pSM DNA site structure required for the assembly of the fully active ParB_pSM-parS_pSM PC in the pSM19035 ParABS system. We propose one possible scenario to explain our findings. However, the model presented here is perhaps not the only possible explanation. Further studies are needed to support or refute the model to advance our mechanistic understanding of this family of partition systems. Direct examination of the effects of the PC torsional strain would be needed to test our current model. Further refinement of the cell-free partition reaction system is hoped to greatly assist our mechanistic understanding of the rich variations of the prokaryotic chromosome/plasmid partition systems.

Newly generated materials from this study are available by request to the corresponding author, Kiyoshi Mizuuchi (kiyoshimi@niddk.nih.gov) until the lab stocks become exhausted or the lab group operation becomes terminated.

Non-fluorescent ParA_pSM-His₆ (wild-type and mutants) was purified as previously described (Pratto et al., 2008). ParA_pSM-GFP-His₆ and ParA_pSM^D60E-GFP-His₆ were purified as described for ParA_F-GFP-His₆ (Vecchiarelli et al., 2010). ParB_pSM wild-type and ParB_pSM-cys, which had three residues (-GCE) added at the C-terminal, were purified essentially as previously described with the addition of reducing agent (2 mM DTT) in the buffers 50 mM Tris–HCl pH 7.5, 100 mM NaCl plus 5% glycerol. Protein concentrations were estimated based on OD₂₈₀ and aromatic amino acid content.

Fluorescence labeling of ParB_pSM was done as described for Alexa647-ParB_F (Vecchiarelli et al., 2013). ParB_pSM-GCE, in 50 mM Tris–HCl pH 7.4, 100 mM NaCl, 0.1 mM EDTA, 10% glycerol, was mixed with Alexa647 maleimide (Invitrogen) at a protein-to-dye ratio of 1:1 and then incubated for 1 hr at room temperature in the dark. Then, 20 mM DTT was added to stop the reaction. Free dye was removed by spin gel-filtration in a G-50 column. The dye labeling efficiency was determined by spectrophotometry to be ~15%, and the labeled protein was mixed with unlabeled protein to prepare 1%-labeled ParB_pSM-Alexa647 used for the experiments. In vivo, ParB_pSM-GCE in combination with ParA_pSM-GFP was fully competent for partition (Maria Moreno-del Álamo, personal communication), and in vitro, dye labeling did not affect its activities, as measured by its ability to stimulate ParA_pSM ATPase and by its parS_pSM DNA binding activity (A. Volante, unpublished results).

The N-terminal peptides of ParB_pSM (residues 1–27) and ParB_P1 (residues 1–30) used in the ATPase assays were synthesized by GenScript. The sequence of ParB_pSM^1-27 and its variant ParB_pSM^{1-27 K10A} were NH₂-MIVGNLGAQKAKRNDTPISAKKDIMGD-CO₂H (≥97% purity) and NH₂-MIVGNLGAQAAKRNDTPISAKKDIMGD-CO₂H (≥96% purity), respectively. The sequence of ParB_P1^1-30 was NH₂-MSKKNRPTIGRTLNPSILSGFDSSSASGDR-CO₂H (≥97% purity).

Two strands of each double-stranded DNA containing heptad repeat (5′-WATCACW-3′) or non-consensus scrambled sequences were synthetized, annealed, and purified by ITD (Integrated DNA Technologies). The forward sequences of a DNA duplex are listed in the following table:

Plasmids encoding wild-type ParA_pSM (Delta, pCB746), mutant ParA_pSM^D60A (pCB755), and ParB_pSM (Omega, pT712ω) have been described previously (Welfe et al. 2005, Pratto et al., 2008; Volante and Alonso, 2015). The pET11a harboring the sequence of his-ParA_pSM^D60E (pET11a-ParA_pSM^D60E) his-ParA_pSM-eGFP (pET11a-ParA_pSM-eGFP) and its variants (D60A, pET11a-ParA_pSM^D60A-eGFP and D60E, pET11a-ParA_pSM^D60E-eGFP) were constructed by GenScript. The ParB_pSM-GCE [72G, 73C, 74E] allele was created by site-directed mutagenesis and then cloned into the ParB_pSM expression plasmid (pT712) to generate pCB1033.

Unless stated otherwise, all ATPase assays were performed as follows: the reaction contained 2 µM ParA_pSM, the indicated concentration of full-length ParB_pSM (or ParB_pSM^1-27), nsDNA (plasmid pBR322 DNA, 60 µM in bp), and parS_pSM DNA duplex in 50 mM Tris–HCl (pH 7.4), 100 mM KCl, 2 mM MgCl₂, and 1.5 mM [γ-³²P]ATP. Labeled ATP was purified before use as previously described (Vecchiarelli et al., 2010). Reactions were incubated at 37°C for 3 hr and analyzed by TLC as previously described (Pratto et al., 2008). The data points shown are the means with error bars (standard errors of mean) of repeat experiments (N) as indicated for each figure panel. N* indicates repeat number for majority of ParB_pSM concentration points as detailed in the Source Data file. Datasets of repeated measurements were fit after subtraction of background measured without ParA_pSM to a modified Hill equation: v – v₀ = (v_max [B]ⁿ)/(K_Aⁿ + [B]ⁿ), and the fit parameters and symmetrized error ranges reflecting the larger error of the 95% confidence interval below and above the mean were estimated using Prism 9 (GraphPad). For [B], total concentration of ParB_pSM was used instead of free ParB_pSM concentration due to technical reasons.

The flow cells coated with lipid bilayer with attached biotin (DOPC plus DOPE-biotin [1%]) were prepared essentially as described in Han and Mizuuchi, 2010 and rinsed with a buffer containing 25 mM Tris–HCl pH 7.4, 150 mM NaCl, and 5 mM MgCl₂ and 0.1 mM CaCl₂. Sonicated and biotinylated DNA was prepared as follows: 250 μl of 10 mg/ml sonicated salmon sperm DNA (Sigma) was sonicated for an additional 5 min (Misonix sonicator 3000, output level 6, pulsed on/off for 10 s each at 16°C) to size-weighted average length of ~500 bp. In order to biotinylate the DNA ends, the sonicated DNA (1 mg/ml) was incubated with 40 μM biotin-17-dCTP (Invitrogen) and 0.6 units TdT (NEB) in the buffer specified by the enzyme manufacturer at 37°C for 30 min. The reaction was stopped by heating at 70°C for 10 min, and unincorporated biotin-17-dCTP was removed by using S-200 HR Microspin columns (GE Healthcare). The DNA was ethanol precipitated and resuspended in TE buffer. To coat the flow cell with sonicated DNA, the DNA prepared as above was dissolved to 1 mg/ml in 25 mM Tris–HCl pH 7.4, 150 mM NaCl, 5 mM MgCl_2, and 0.1 mM CaCl₂, infused into the assembled flow cell, and incubated overnight at 4°C. Unbound DNA was removed by rinsing with 50 mM Tris–HCl (pH 7.4), 100 mM KCl, 2 mM MgCl₂, and 10% glycerol.

Total internal reflection fluorescence (TIRF) illumination and microscopy, as well as the camera settings, were essentially as described (Hwang et al., 2013). Combined beam of a 488 nm diode-pumped, solid-state laser (Coherent) and a 633 nm HeNe laser (Research Electro-Optics) was pointed through a fused silica prism onto the top side of the sample flow cell. Fluorescence emission was collected through a ×40 Plan Apo VC, NA 1.4 oil-immersion objectives (Nikon) and magnifier setting at ×1.5. The laser excitation lines were blocked below objective lens with notch filters (NF03-488E and NF03-633E, Semrock). The fluorescence images were captured by an EMCCD camera (Andor IXON +897) through a Dual-View module (630DCXR cube, Photometrics; short/long pass filters, SP01-633RS, LP02-633RE, Semrock). Microscopy experiments were carried out at room temperature (~23°C). Typical camera settings were digitization 16 bit at 1 MHz, preamplifier gain 5.2, vertical shift speed 2 MHz, vertical clock range: normal, EM gain 40, EMCCD temperature set at –98°C, baseline clamp ON. Images were acquired at exposure time 100 ms with frame rate 1, 0.4, or 0.2 Hz using MetaMorph 7 software (Molecular Devices) and analyzed using ImageJ software (NIH) as described (Hwang et al., 2013). Data were analyzed with Prism 9 (GraphPad).

ParA_pSM and ParB_pSM density on the nsDNA-carpet (dimers/μm²) were calculated from the fluorescence intensity of acquired images essentially according to the procedure described in Figure S4 legend in Vecchiarelli et al., 2016. The labeled protein samples used in the experiments were diluted to different concentrations (0–8 μM) in a buffer (50 mM Tris–HCl, pH 7.4, 100 mM KCl, 2 mM MgCl₂, 10% glycerol, 1 mM DTT, 1 mg/ml α-casein, and 0.6 mg/ml ascorbic acid) and infused into flow cell coated with 1,2-dioleoyl-sn-glycero-3-phosphocholine. Fluorescence intensity data were collected before the protein arrival (background), with the protein sample in the flow cell, and after washing the flow cell with buffer (background) using the illumination and image acquisition parameter settings used for the series of experiments for which the conversion parameters were prepared. The fluorescence signal from the protein sample in the solution was obtained by subtracting the background signal with buffer only (mostly camera dark noise). From the wavelength, the refractive indices of fused silica slide glass and the reaction buffer, and the illumination angle, the evanescence penetration depths were calculated to be 131 and 170 nm for 488 and 633 nm excitation beams, respectively. The protein concentration and evanescence penetration volume yielded the number of protein molecules, and when bound to the flow cell surface, would produce the observed fluorescence signal, yielding the conversion factor used.

Two inlet flow cells were assembled and coated with sonicated salmon sperm DNA as described (Vecchiarelli et al., 2013). One inlet was connected to syringe A containing ‘binding solution’ and the other to syringe B containing ‘wash solution.’ For protein binding to the nsDNA-carpet, syringe A delivered at 5 μl/min, while syringe B at 0.5 μl/min. To start the wash, the flow rates of the two syringes were reversed keeping the total flow rate constant. Experiments were done at room temperature in pSM19035 buffer: 50 mM Tris–HCl (pH 7.4), 100 mM KCl, 2 mM MgCl₂, 10% glycerol, 1 mM DTT, 0.1 mg/ml α-casein, 0.6 mg/ml ascorbic acid, and 1 mM ATP. Additional components of the binding and washing solutions were as specified in the figure legends. Syringe components were preincubated before infusion at 23°C for 30 min or longer. The association and dissociation data points represent the means with error bars (standard errors of mean) of repeat experiments (N) as indicated for each panel. The dissociation curves were fitted to a single- or a double-exponential equation after subtraction of background in the absence of protein with plateau level set to zero when necessary to avoid large plateau values, and the fraction of fast-dissociating population, the k_off, and symmetrized error ranges reflecting the larger error of the 95% confidence interval below and above the mean were estimated using Prism 9 (GraphPad).

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We are grateful to our colleagues Keir Neuman, Barbara Funnell, Anthony Vecchiarelli, James Taylor, Michiyo Mizuuchi, Min Li, Masaki Osawa, and William Carlquist for helpful suggestions and discussion, and to Maria Moreno-del Álamo for providing the information concerning the in vivo functionality of ParA_pSM-GFP/ParB_pSM-GCE. This work was supported by the intramural research fund for National Institute of Diabetes and Digestive and Kidney Diseases (KM) and the Ministerio de Ciencia e Innovación, Agencia Estatal de Investigación (MCIN/AEI)/FEDER, 2018-097054-B-I00 and 2021AEP031 to JCA.

This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.