bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license.

1 A System with Tunable Copy Number

2 Miles V. Rouches and Yasu Xu 3 Field of Biophysics, Cornell University, Ithaca, NY 14853, USA

∗ 4 Louis Cortes and Guillaume Lambert 5 School of Applied and Engineering Physics, Cornell University, Ithaca, NY 14853, USA 6 (Dated: July 13, 2021) are one of the most commonly used and time-tested molecular biology platforms for genetic engineering and recombinant gene expression in . Despite their ubiquity, little consideration is given to metabolic effects and fitness costs of plasmid copy numbers on engineered genetic systems. Here, we introduce two systems that allow for the finely-tuned control of : a plasmid with an anhydrotetracycline-controlled copy number, and a massively parallel assay that is used to generate a continuous spectrum of ColE1-based copy number variants. Using these systems, we investigate the effects of plasmid copy number on cellular growth rates, gene expression, biosynthesis, and genetic circuit performance. We perform single-cell timelapse measurements to characterize plasmid loss, runaway plasmid replication, and quantify the impact of plasmid copy number on the variability of gene expression. Using our massively parallel assay, we find that each plasmid imposes a 0.063% linear metabolic burden on their hosts, hinting at a simple relationship between metabolic burdens and plasmid DNA synthesis. Our plasmid system with tunable copy number should allow for a precise control of gene expression and highlight the importance of tuning plasmid copy number as tool for the optimization of synthetic biological systems.

7 INTRODUCTION 37 based plasmids to study the impact of copy number on 38 cellular growth and metabolism. Specifically, we first in-

39 8 Plasmids are an invaluable tool in biotechnology and troduce a strategy for finely tuning plasmid copy number 40 9 genetic engineering due to their small size and the ease in a way that is dependent on the concentration of an ex- 41 10 with which they can be engineered for biotechnological ogenous inducer molecule. Then, we develop a method 42 11 applications. Recent studies have highlighted the poten- which uses a massively parallelized assay for rapid screen- 43 12 tial benefits of fine-tuning plasmid copy number to the ing of the dependence of a system on plasmid copy num- 44 13 task at hand. For example, runaway replication has been ber. 14 used as a method for increasing cellular protein produc- 15 tion [1, 2]. Recent works have also leveraged plasmid 16 copy numbers to amplify a signal through a large increase 17 in copy number [3, 4] and to increase the cooperativity 18 and robustness of synthetic gene circuits [5]. Single cell 19 measurements have also been used to accurately quantify 20 the prevalence of plasmid loss within a population [6]. 21 While recent studies have probed the effect of plas- 22 mid copy number on host cell growth [7, 8] and the role 45 We further characterize our systems at both the macro- 23 of plasmids on central metabolic processes [9], we still 46 scopic and single cell level to uncover a relationship be- 24 lack a quantitative description of the relationship be- 47 tween gene copy number, protein production, and cellular 25 tween plasmid copy number and host metabolic burden. 48 growth rates. Our approach also leads to insights into the 26 Indeed, recombinant gene expression results in signifi- 49 variation in plasmid copy numbers and the phenomena 27 cant metabolic burden on bacterial cells [10–12] which 50 of plasmid loss and runaway replication at the single- 28 may interfere with normal cellular processes and even 51 cell level. We finally demonstrate that manipulations of 29 lead to cessation of growth. The control of plasmid copy 52 plasmid copy numbers can be used to tune the behav- 30 number can also drastically alter the behavior of simple 53 ior of synthetic gene circuits such as a simple CRISPRi 31 repression systems [13, 14] which may significantly im- 54 genetic inverter, and control the expression of complex 32 pact engineered gene circuits and should be important 55 biosynthetic pathways to optimize the production of the 33 considerations for plasmid-based biotechnology and bio- 56 pigment Violacein. Our work thus provides a straightfor- 34 engineering applications 57 ward method for generating copy number variants from 35 In this work, we use two novel approaches to actively 58 a specific plasmid backbone, which we anticipate will ex- 36 control the copy number of pUC19 and other ColE1- 59 pand the synthetic biology toolkit and constitute the ba- 60 sis for more nuanced attention to plasmid copy number 61 in synthetic gene circuits, protein biosynthesis, and other ∗ Corresponding author 62 biotechnological applications. bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 2

A pUC19 Plasmid map D E 0.03 Inhibition window 100 Pbla Pp,kp (n steps) Plac RNA-p 0.02 +1 +111 +360 +555 MCS pUC-pTet 1 10 Ptet 0.01 ampR ORI LacZa OD 600 P ,k 0 RNA-i i i 0 300 600 Time (mins) Growth rate (1/min)

Plasmid Copy Number 1 B RNA-p C 0 0.01 0.1 1 10 100 1 10 100 aTc Concentration (ng/ml) Copy number RNA-i 100 1000 RNA-i/p 10 PCN 1 F pUC-pLac G 100 0 60 120 Division time 0.020 ρ 300 pUC19 copy number Plac 10 0.015 RNA-i/p RNA-p hyperbolic (n=1) binding priming 200 Runaway replication 1 stableexponential plasmid copy (n=∞) number 0.010 Plasmid Copy Number

1 100 OD 600 0 0.005 0 300 600 Plasmid loss Time (mins) Growth Rate (1/min)

Plasmid Copy Number 30 RNA-p 0.00 0.25 0.50 0.75 1.00 1.25 -4 -2 0 0.000 release 0 10 10 10 30 100 200 300 No Plasmid RNA production ratio IPTG Concentration (mM) Copy Number replication replication ( )

Fig. 1. Replication of ColE1 Origin Plasmids. A) Map of the pUC19 plasmid. The consists of a 555 bp region containing the priming RNA (RNA-p), the inhibitory RNA (RNA-i), the inhibition window, and the site of replication initiation (ORI). B) Diagram of the steps leading to ColE1 replication. In the ColE1 Origin, reversible binding of the inhibitory RNA-i to RNA-p exercises control over the copy number by inhibiting replication (left branch). Production of RNA-p transcript may result hybridization with the ORI, which leads to plasmid replication (right branch). C) Predicted plasmid copy number for hyperbolic (upper) and exponential (lower) initiation models [15] plotted as a function of RNA-p transcription initiation rate. Inset shows the predicted dependence of plasmid copy number on host cell division time. Model parameters:  = 0.545 / min, ρ = 0.65, r = 0.01 / min, ki = 0.95 / min. D) Plasmid copy number measurements made by digital droplet PCR of the aTc- inducible pUC-pTet plasmid at increasing aTc concentrations. E) Growth rates of cells hosting the pUC-pTet plasmid at increasing plasmid copy numbers. We see a marginal effect of plasmid copy number on cellular growth rates, inset shows individual growth curves colored by aTc concentration. F) Plasmid copy number measurements of a plasmid which contains an additional copy of the inhibitory RNA under the control of an IPTG-inducible promoter. G) Growth rate vs plasmid copy number of the IPTG-inducible plasmid shows a strong metabolic burden at high levels of RNA-i expression.

63 RESULTS 84 division. 85 Specifically, as the ratio of RNA-p transcription rate 86 kp to RNA-i transcription rate ki is increased, plasmid 64 Model of ColE1 Plasmid Replication and Copy 87 copy number N will also increase according to 65 Number p

" 1 #n 66 The molecular mechanisms that regulate plasmid copy  + r ρk  n N = p − 1 (1) 67 number control have been well characterized [16]. In p ki − kp r 68 the ColE1 origin of replication, used extensively in this 69 work, two RNAs control replication (Fig. 1A): the prim- 88 where  is the RNA degradation rate, ρ is the RNA-p 70 ing RNA (RNA-p), which acts in cis as a primer near 89 priming probability, r is the cellular growth rate, and 71 the origin of replication, and the inhibitory RNA (RNA- 90 n is the number of rate-limiting steps in the inhibition 72 i) which is transcribed antisense to RNA-p and acts in 91 window [15]. The n → 1 and n → ∞ limits represent 73 trans to inhibit replication through binding to the prim- 92 hyperbolic and exponential inhibition, respectively (Fig. 74 ing RNA prior to hybridization near the origin (Fig. 1B) 93 1C). 75 [17]. 94 As the priming:inhibitory RNA production ratio kp/ki 76 A simple macroscopic-scale model developed by Pauls- 95 approaches 1, the number of inhibitory RNA will not 77 son et al. recapitulates these observations and predicts 96 be present in high enough number to sequester priming 78 the functional dependence of plasmid copy number on the 97 RNA transcripts, thereby failing to limit plasmids from 79 molecular details of this process [15] (Fig. 1C). This pro- 98 duplicating in a process called runaway replication (Fig. 80 cess is classified as inhibitor-dilution copy number control 99 1C). In addition, increasing the plasmid copy number 81 [18], owing to the fact that the copy number is controlled 100 may increase the metabolic burden associated with plas- 82 through a combination of replication inhibition and dilu- 101 mid expression and reduce the growth rate r, which in 83 tion of plasmids and regulatory components due to cell 102 turn increases the plasmid copy number further as r → 0. bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 3

pUC-pTet-sfGFP 129 ber with aTc induction (Fig. 1D). Copy numbers ranged A Ptet B 0.025 ~50 plasmids 130 from 1.4 copies per cell in the complete absence of aTc to sfGFP per cell −1 Pconst 131 roughly 50 copies per cell at full induction (100 ng mL

~1-3 plasmids 0.020 132 aTc). We saw a very similar relation in the plasmid yield per cell 133 from overnight cultures grown at varying aTc concentra-

134 [aTc] tions (Sup.).

PCN Growth Rate (1/min) fitness 0.015 0 0.01 0.1 1 10 100 135 We subsequently measured the growth rate of cells har- Low protein expression High protein expression Risk of plasmid loss Risk of runaway replication aTc Concetration (ng/mL) 136 boring the pUC-pTet plasmid as a function of the aTc 137 concentration in a microplate reader (Fig. 1E). Though C D 138 the absolute difference in growth rates is small, a cell with 103 103 139 a few copies of the plasmid grows roughly 10 % faster 140 than a cell with 50 plasmids, we can clearly see that in- 141 creasing the plasmid copy number has a weak, though GFP (AU) 102 GFP (AU) 102 142 consistent, effect on the host cell growth rate.

143 To complement the work above, we sought to demon- 0 0.01 0.1 1 10 100 0 20 40 aTc conc (ng/mL) Plasmid Copy Number 144 strate that the transcription level of inhibitory RNA can 145 be used to control ColE1 copy number as well. Seeking to 146 do this with minimal perturbation to the native pUC19 Fig. 2. Control of Gene Expression through Induction of Plas- 147 origin of replication, we inserted a second copy of the mid Copy Number. A) Schematic of the effects of adding aTc 148 inhibitory RNA downstream of the IPTG-inducible Lac to the pUC-pTet-GFP plasmid. As the concentration is in- 149 promoter on the pUC19 plasmid. This gives us control creased the number of plasmids and GFP genes increases as 150 over the inhibition of replication through the addition of well as the metabolic cost of the plasmid. B) Growth rates 151 exogenous IPTG (Fig. 1F). of cells hosting the pUC-pTet-sfGFP plasmid as a function of aTc concentration. C) GFP Production of the pUC-pTet- 152 When no IPTG was added to the cells we measured GFP plasmid grown at different aTc concentrations. D) GFP 153 ∼270 plasmids/genome, a number that agrees with what production of the pUC-pTet-GFP plasmid as a function of 154 we have measured for the standard pUC19 plasmid, indi- plasmid copy number as measured by digital droplet PCR. 155 cating that our system efficiently represses the additional 156 inhibitory RNA transcript and does not affect the copy 157 number in the absence of induction. Induction of the in- 103 On the other hand, as the RNA production ratio kp/ki 158 hibitory RNA greatly reduced plasmid copy number to 104 decreases, the number of plasmids inside each cell will ap- 159 approximately 30 plasmids/genome at full RNA-i induc- 105 proach 1 and small variations in the number of plasmids 160 tion, demonstrating that tuning RNA-i production rate 106 apportioned to each daughter cell during cell divisions 161 can be used to control the plasmid copy number. 107 may lead to the catastrophic consequence of plasmid loss. 162 It is interesting to note that we observed paradoxi- 108 These two stresses, plasmid loss and metabolic burden, 163 cal effects on the growth rate from cells harboring this 109 necessitate the evolution of copy number control systems 164 plasmid - at high plasmid copy numbers we see a faster 110 to ensure that neither stress results in the loss of the 165 growth rate and at lower copy numbers we see a markedly 111 plasmid [19]. 166 slower growth rate (Fig. 1G). These observations are 167 markedly different from what we see in the pUC-pTet 168 plasmids, where increased plasmid copy number has a rel-

112 Tuning Priming and Inhibitory RNA Transcription 169 atively modest effect on the growth rate of host cells. In- 113 to Control ColE1 Copy Number 170 stead, the reduction in the maximum cell density reached 171 by the IPTG-induced populations (Fig. 1G, inset) sug- 172 gests that overproduction of the inhibitory RNA may 114 The Paulsson model of ColE1-type replication predicts 173 completely abolish plasmid replication in some cells, pre- 115 that the RNA-p transcription initiation rate should de- 174 venting them from dividing further. 116 termine the plasmid copy number without drastically al- 117 tering the sensitivity of copy number control [18]. Thus, 118 in order to only control the activity of the promoter up- 175 119 stream of the priming RNA, we first replaced the native Gene Expression Scales with Plasmid Copy Number 120 priming promoter in the pUC19 plasmid with an anhy-

121 drotetracycline (aTc) inducible promoter (Fig. 1D). Liq- 176 Understanding the interplay between plasmid copy 122 uid cultures of E. coli overexpressing TetR and carrying 177 number, host cell growth rate, and gene expression is 123 this plasmid were grown in aTc concentrations spanning 178 key to optimizing protein production in bacterial cell cul- 124 4 orders of magnitude and the plasmid copy number was 179 tures. If too much protein is produced, cells may stop 125 determined via digital droplet PCR (ddPCR) from total 180 growing or be out competed by cells harboring plasmids 126 DNA isolates by measuring the ratio of the bla gene on 181 that lack the correct insert. If protein production is too 127 the plasmid to the genomic dxs gene [20]. 182 low, on the other hand, one may have to process massive 128 This plasmid showed robust induction of copy num- 183 amounts of cells to obtain a usable amount of their target bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 4

A B C 0 D 0.1 ng/mL 10 1.00 0.3 ng/mL 1 ng/mL 100 ng/mL N=59,757 0.75 <µ1> = 0.0148 10−1 1.0 10 ng/mL 0.50 N=81,511 0.5 FC 0.0 <µ > = 0.0157 Fold-change 0.25 0.3 3 ng/mL −2 0 1 100 10 aTc N=71,885 <µ > = 0.0172 Rel. Prob. Density 0.1 0.00 0 0.01 0.1 1 10 100 −2.25 −2 −1.75 −1.5 1 ng/mL 10 10 10 10 N=35,277 aTc concentration (ng/mL) Growth Rate (1/min)

0.3 ng/mL N=33,156 E F 60% 100% 100 3 0.1 ng/mL Run- 10 100 N=13,800 10 away 40% repli- 1 0.01 ng/mL 3 cation 1 0.1 N=10,276 10% 0.01 0.3 20% 0.1 Plasmid 0 ng/mL loss 0.3 N=1,455 0 0 0.01 Viable Cells Fraction 0% 10 2 10 4 Elongated Cells Fraction 0 500 1000 1500 0 500 1000 1500 Intensity Time (mins) Time (mins)

Fig. 3. Microscopic Analysis of sfGFP Expressed by the pUC-pTet. A) Representative snapshots of microfluidic chambers at various aTc concentrations. Slow-growing cells at high copy number are indicated by orange arrows and cells at 0.3 ng mL−1 of aTc in an “activated” state are indicated by purple arrows. B) Distribution of cellular fluorescence intensities at increasing aTc concentrations. C) Normalized mean fluorescence as a function aTc concentration, log scale. Inset: Normalized mean fluorescence vs aTc concentration on a linear scale. D) Growth rate distributions of cells growth at 0.1, 0.3 and 1 ng mL−1 aTc. E) Fraction of slow-growing, elongated cells over time at several aTc concentration. This measurement is indicative of the extent to which cells are burdened by the high plasmid copy numbers. F) Proportion of viable cells as a function of time at various aTc concentrations. The sharp decrease in the number of viable cells at low aTc concentrations provides insight into the extent to which plasmid loss occurs at low copy numbers.

184 protein. Understanding the interplay between plasmid 212 concentration (Fig. 2C). In addition, Fig. 2D shows a 185 regulation and protein expression is key to solve protein 213 nearly linear increase in the fluorescence with the plasmid 186 production maximization problems. 214 copy number, suggesting that the metabolic costs asso-

215 187 To directly measure the effects of plasmid and gene ciate with protein overproduction only marginally impact 216 188 copy number on gene expression and host cell growth the overall GFP production rate. Our pUC-pTet plasmid 217 189 rate, we first inserted sfGFP downstream of a consti- provides a simple way to control gene expression without 218 190 tutive promoter on our pUC-pTet plasmid, whose copy the need to insert an inducible promoter upstream of the 219 191 number is controlled through exogenous aTc induction. gene of interest. 192 This gives us a straightforward way of monitoring the ef- 193 fects of plasmid copy number on gene expression. Cell

194 cultures with this tunable copy number plasmid were 220 Metabolic Costs of the pUC-pTet Plasmid at the 195 grown in a microplate reader where simultaneous mea- 221 Single-Cell Level 196 surements of fluorescence and growth rate could be made.

197 We first observe clear differences in growth rates as 222 While macroscopic measurements show a clear interde- 198 a function of both plasmid copy number and aTc con- 223 pendence between gene expression and copy number, sin- 199 centration (Fig. 2B). We see a roughly 15 % decrease in 224 gle cell measurements can provide a more powerful path 200 host growth rate at maximal induction of copy number 225 towards understanding the relationship between plasmid 201 (∼ 50 copies per cell). This metabolic burden is a sig- 226 copy number and gene expression. In particular, single- 202 nificantly higher than the pUC-pTet plasmid (Fig. 1E), 227 cell studies allow us to determine the distributions of gene 203 indicating that the majority of the reduction in growth 228 expression of single cells at varying plasmid copy num- 204 rate comes from the overexpression of GFP and not from 229 bers and directly observe the phenomena of plasmid loss 205 the increased plasmid copy number. In this sense, addi- 230 and runaway replication.

206 tion of GFP to the plasmid provides a stronger coupling 231 To first corroborate protein production measurements 207 between the host growth rate and plasmid copy number 232 with plasmid copy number, we observed the pUC-pTet- 208 by taking up additional cellular resources. 233 sfGFP system for up to 150 generations in a microfuidic

209 Despite the fitness costs associated with increasing the 234 device [21, 22]. 210 plasmid copy number, we nevertheless observe an almost 235 In agreement with population-wide measurements 211 perfect GFP induction curve when compared to the aTc 236 (Fig. 2C), individual cells show an increase in the mean bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 5

A 1) Library preparation: 2) Transformation: 3) Growth and selection: 4) Sequencing Pp Pi >103 and counting: Linearization .

. n . f n . . plasmids lig . + KLD assembly 0.25x Electroporation 0.25x n3 Priming promoter Pp library -35 -10 |+1 OD 0.25x pUC19 : TTGAGAtcct...cgcgTAATCTgctgctT n2 Pp : TTKMMMTcct...cgcgTWWRMTgctgctN E. coli Fold-change: log2(n1/nlig) time Growth rate: (ni+1/ni )^0.25 Inhibiting promoter Pi library >300k transformants Recovery n1 -35 -10 |+1 (1h) Plasmid copy number: pUC19 : TTGAAGtggt...cggcTACACTagaagA n1/nlig exp(-t/Tau) Pi : TTKMMMtggt...cggcTWWRMTagaagN

B C D 800

830 364 300 Pp Library Pi Library Pp 100 Pi 30 ddPCR n=830 n=364 10 3

1

1 3 10 30 100 300 800 Sequencing Measurement E 0.030 metabolic cost = 0.063% per 0.025 plasmid Ranked promoter list (Priming promoter) Ranked promoter list (Inhibiting promoter) 0.020 1 1

Growth Rate 0.015 Linear Metabolic Burden

Growth Rate (1/min) Error > 25% 5% < Error < 25% 0.010 Error < 5%

−2 0 2 −2 0 2 1 3 10 30 log10(PCN) log10(PCN) 0.1 100 300 800 Plasmid Copy Number

Fig. 4. Massively Parallelized Copy Number Assay. A) Procedure for massively parallelized copy number assay. The promoter upstream of the priming RNA was randomized with site directed mutagenesis. 2. Plasmids were ligated and cells were transformed with this library of plasmids with high efficiency and grown until OD600 of 1.0, 3. at which point a fraction of the culture was diluted and regrown and the remaining fraction was saved for plasmid extraction. 4. Sequencing libraries were generated from isolated plasmids and sequenced using next generation sequencing. Growth rates and Copy numbers were then calculated from the frequencies of sequencing counts present at each time point. Digital PCR was performed to measure the copy number of individual constructs so that absolute copy numbers could be estimated from sequencing measurements. B) Ranked list of plasmids generated through mutations to the promoter driving the priming RNA in this work (n = 830). (lower) The distribution of copy numbers arising from mutations to the priming RNA promoter. C) Ranked list of plasmids generated through mutations to the promoter driving the inhibitory RNA in this work (n = 364). D) Agreement between plasmid copy number as measured by sequencing counts and digital droplet PCR. E) The relationship between plasmid copy number and growth rate as measured across 830 variants of the pUC19 plasmid. The size of each data point is inversely proportional to the error in the measurement, with the top, middle two and lowest quartiles representing an error of less than 5 %, between 5 % and 25 %, and more than 25 %, respectively. Data from Fig. 1E is plotted (red to blue gradient points) A linear metabolic burden seems to account for the observed relation reduction in growth rate with increasing copy number.

237 fluorescence with plasmid copy number (Fig. 3A). We 249 intensity (Fig. 3C) is in good agreement with population 238 also observe a marked increase in the fluorescence of some 250 wide measurements (Fig. 2C), but each population sees a 239 cells at high aTc concentrations. Interestingly, cells with 251 steady increase in the number of individual cells with in- 240 an increased fluorescence level are found to be elongated 252 creased GFP levels (Fig. 3A, red arrows). As the plasmid 241 (Fig. 3A and Fig. S2). The intensity distribution (Fig. 253 copy number increases further, a growing number of cells 242 3B) undergoes a profound change as the plasmid copy 254 reach an “activated” state where they show fluorescence 243 number increases: the distribution is peaked at a very 255 levels as high as those full pTet induction. Eventually, all −1 244 low intensity at low plasmid copy number but, as the 256 cells for aTc concentrations above 3 ng mL reach this 245 copy number increases, the distribution widens and a 257 activated, fully-induced state (Fig. 3C).

246 high-intensity tail emerges. 258 As the plasmid copy number sharply increases between −1 −1 247 The most drastic shift in the distribution occurs be- 259 0.1 ng mL and 1 ng mL of aTc, we observe a clear −1 −1 248 tween 0.3 ng mL and 1 ng mL , where the average cell 260 drop in average growth rate (Fig. 3D and Fig. S4) similar bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 6

261 to bulk measurements (Fig. 2B). The distribution of cel- 316 tion with the fold change in sequencing counts from the 262 lular growth rates (Fig. 3D) show that this drop is largely 317 library prior to transformation, we determined the rela- 263 due to a fraction of cells that suffer from reduced growth 318 tive copy numbers of every plasmid in this library (Figs. 264 rate. This is likely due to the metabolic burden of plas- 319 4B-C) and their growth rates (Fig. 4D). Our method 265 mid maintenance and GFP expression severely limiting 320 was able to generate 1,194 plasmids, each with a distinct 266 the growth of a subset of the cellular population. Above 321 plasmid copy number ranging from less than 1 copy per −1 267 1 ng mL aTc, the growth rate of healthy cells seems 322 cell to close to 800 copies per cell (Table S1).

268 to stabilize (Fig. S4). However, we notice an increasing 323 To make sure our sequencing-based method yields ac- 269 fraction of elongated cells (Fig. 3E), many of which have 324 curate plasmid copy number measurements, we next 270 lost ability to grow and turned extremely bright (supple- 325 sought out to calibrate our sequencing-based numbers to 271 mentary movie M1). The inability of a fraction of cells to 326 absolute copy numbers made using digital droplet PCR 272 grow may explain why growth rate continually decreases 327 measurements. Individual constructs of 9 promoters were 273 even though plasmid copy number saturates at high aTc 328 selected (light blue datapoints in Fig. 4B) and their ab- 274 levels in bulk measurements (Fig. 2B). 329 solute copy number were determined via digital PCR by 275 At low plasmid copy numbers, we see a decreased num- 330 the ratio of dxs to bla genes [20]. There is a good agree- 276 ber of cell in each microfluidic chamber. In fact, nearly 331 ment between fold change in sequencing abundance and 277 every cells undergoes lysis within the first 1500 min of 332 the plasmid copy numbers as measured by digital droplet −1 278 tracking for aTc concentrations lower than 1 ng mL , 333 PCR (Fig. 4D), indicating that the fold change in abun- 279 suggesting that cells may not be resistant to carbenicillin 334 dance of sequencing counts with a modest correction ap- 280 due to plasmid loss (Fig. 3F). 335 plied to correct for differences in growth rates is a sound

281 These observations highlight the delicate balance be- 336 measure of plasmid copy number. 282 tween plasmid loss at low copy number and heightened 337 The simultaneous measurement of plasmid copy num- 283 metabolic burdens at high copy numbers. Our results 338 bers and the growth rates of their host cells provides 284 demonstrate that both phenomena occur stochastically 339 us with a powerful measurement of the interdependence 285 and on a per-cell basis. 340 of copy number and metabolic burden. While a high 341 plasmid copy number should reduce host cell growth, to 342 what extent that occurs is not known. Our measure-

286 Massively Parallelized Assay Determines Copy 343 ments of plasmid copy numbers and growth rates for the 287 Numbers of ColE1-derived Plasmids 344 830 plasmids in the priming promoter library, all with 345 nearly identical origins of replications, provide a power- 346 ful way to deduce this relationship. 288 Though we were able to obtain robust control over 347 In Fig. 4E, we observe that, as the plasmid copy num- 289 plasmid copy number with our inducible plasmids, we 348 ber increases, the growth rate of host cells decreases. 290 next sought to exert control over plasmid copy number 349 A simple relationship in which the metabolic burden of 291 without the need for external inducer molecules. In order 350 plasmid expression is linear with the copy number largely 292 to survey the landscape of possible plasmid copy numbers 351 agrees with these findings –ie. cost = a · b · x + b, where 293 that arise from changes of the transcription rates of the 352 x is the plasmid copy number, with a = 0.063 % and 294 RNAs controlling ColE1 replication, we simultaneously −1 353 b = 0.023 min . This relationship can arise from a sim- 295 constructed 1024 different variants of the pUC19 plas- 354 ple resource allocation model in which each plasmid takes 296 mid through site directed mutagenesis of the −35 and 355 up a small fraction of the resources necessary for cellu- 297 −10 boxes of the promoter controlling the priming RNA 356 lar growth proportional to the plasmid size (pUC19 is 298 and at the +1 site of the priming RNA transcript (Fig. 357 2686 bp in size, which is approximately equal to 0.058% 299 4A, Pp library). The same procedure was performed on 358 the length of the E. coli genome). This results opens up 300 the promoter controlling the inhibitory RNA (Fig. 1A, 359 the possibility for improved models of plasmid evolution 301 Pi library), giving us 2 separate libraries of plasmids: 360 and replication, as they indicate that plasmids impose a 302 one with a variable priming RNA transcription rate kp 361 metabolic burden that is linear with their copy number 303 and another with a variable inhibitory RNA transcription 362 on their hosts. Knowledge of this relationship between 304 rate ki. 363 host growth and plasmid copy number may shed light on 305 These libraries were constructed using multiplexed 364 the evolution of cis and trans acting regulatory elements 306 PCR and KLD assembly [23] and electroporated into 365 as well as the stringency of plasmid copy number control 307 competent E. Coli cells (Fig. 4A, step 2). Cell cultures 366 mechanisms. 308 were grown in selection conditions until they reached an 309 OD600 of 1.0 and subsequently diluted by a factor of 4 310 and grown again to an OD600 of 1.0 (Fig. 4A, step 3). 311 Excess culture from each step of dilution was set aside 367 Plasmid Copy Number Controls Violacein 368 312 for sequencing (Fig. 4A, step 4). Biosynthesis

313 For each recovered mutant, the abundance of each con- 314 struct at each time point was used to infer the growth rate 369 We sought to demonstrate the potential for using plas- 315 of each individual construct. By coupling this informa- 370 mid copy number to control complex biosynthetic pro- bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 7

VioD 388 and observed a clear increase in the violacein production A VioB O VioE VioC O O O O 389 with increasing aTc concentrations (Fig. 5C). N N N O O N N VioA VioB N VioE N N VioD VioC O N N 390 Additionally, we inserted the VioABCDE pathway into N O N O O N VioA O O VioC N 391 a pUC plasmid under the control of a moderate-strength Tryptophan VioB Violacein 392 promoter, and we used site directed mutagenesis of the 393 origin of replication, in the manner described above, to

394 B VioABCDE pathway D VioABCDE pathway generate a plasmid copy number library (Fig. 5D). Af- 395 ter this library was transformed into competent cells, we VioA VioB VioC VioD VioE 396 Gibson assembly Pp library picked individual colonies from a plate and sequenced 7.754kb insert (1024 sequences) 397 their plasmids in order to infer their copy number from

ampR 398 the Pp library data (Table S1). We also measured the vi- 399 olacein production from cell extracts and found a clearly pTet KLD assembly 400 positive relation between the plasmid copy number and ampR >10k transformants 401 High PCN Violacein production (Fig. 5E). Interestingly, we ob- Low PCN High PCN Low PCN variants 402 served a marked decrease in violacein yields as the plas- [aTc] variants 403 mid copy number is increased above 500 copies per cell, Violacein production Violacein production 404 and growth curve measurements performed using a plate C E 405 reader show that the decreased yield can be attributed 100 406 to a longer lag time (Fig. 5E, inset). Our results show 1.00 400 407 that using tunable copy number thus constitutes a gen- 0.75 200 408 eralizable method for the scaling-up and optimization of

−1 Lag Time (mins) 0.5 1.0 10 0.50 Viol. Yield 409 complex biosynthetic pathways.

0.25 Violacein Yield (AU) Violacein Yield (AU) 10−2 0.00 410 0 0.01 0.1 1 10 100 5 Control of a CRISPRi system through sponge 50 aTc concentration (ng/mL) 100 300 500 411 binding sites Plasmid Copy Number

Fig. 5. Optimization of Violacein Production by Tuning Plas- 412 Controlling the plasmid copy number may also be use- mid Copy Number. B) The complete VioABCDE pathway 413 ful in synthetic biology and in the creation of robust was inserted in the pUC-pTet plasmid. C) Controlling vio- 414 gene circuits [3, 5, 27, 28]. To test this, we used a lacein biosynthesis by increasing aTc concentrations. D) A 415 simple CRISPR-dCas12a inverter to demonstrate the ef- library of variable copy number plasmids was created by per- 416 fect of adding “sponge” binding sites to a simple genetic forming mutagenesis on the priming promoter of a ColE1- 417 circuit through manipulation of plasmid copy numbers. based plasmid containing the VioABCDE pathway. E) Viola- 418 Sponge sites are additional binding sites used to titrate cein production peaks as the plasmid copy number nears 300, 419 out (“soak up”) transcription factors, and they have been with a marked decrease in the yield as plasmid copy number is increased beyond 500. Inset: relationship between the vio- 420 used to redirect flux in metabolic pathways to optimize lacein yield and lag time of the population when grown in a 421 arginine production [29], to activate silent biosynthetic plate reader. 422 gene clusters [30], to tune gene expression timing [31], 423 and mitigate protein toxicity [32]. In these studies, vari- 424 able numbers of sponge sites are placed on high or low 425 copy number plasmids. 371 cesses comprising of several proteins and enzymes. We 426 Here, we added a sponge site to a few isolates from 372 chose to focus on Violacein (Fig. 5A), a purple pigment 427 our tunable copy number library to alter the qualita- 373 formed by the condensation of two tryptophan molecules 428 tive behavior of a simple genetic circuit. Specifically, we 374 [24] which is commonly used to demonstrate biosynthe- 429 use a two plasmid system in which a low-copy plasmid 375 sis optimization [25]. The violacein biosynthetic pathway 430 (pSC101, 2-5 copies per cell) contains an aTc inducible 376 contains five enzymes, VioABCDE, each under the con- 431 CRISPR-Cas12a guide RNA which can bind to an active 377 trol of a different promoter and spanning approximately 432 site Sa and turn OFF a promoter driving sfGFP, while 378 8kb of DNA (Fig. 5B). Hence, we reasoned that tuning 433 the second plasmid contains a decoy binding site Sd site 379 the plasmid copy number of the whole Violacein produc- 434 that sequesters CRISPR-Cas12a molecules away from the 380 tion pathway may serve as an effective method to scale 435 active site Sa. The sponge plasmids were picked out from 381 up the expression of VioABCDE while keeping the pro- 436 our library of variable copy number plasmids generated 382 duction of each enzyme under the same stoichiometric 437 in Fig. 4B and span copy numbers from 30 to 270. 383 ratio. 438 Time-series measurements of the fully-induced 384 To achieve this, constructed a pTet inducible plasmid 439 CRISPRi inverter show that the OFF state of the 385 harboring the Violacein biosynthesis pathway using Gib- 440 inverter has a higher fluorescence level as the number of 386 son assembly [26] (Fig. 5B). We then grew cells with the 441 sponge sites in increased (Fig. 6A). This suggests that 387 pUC-pTet-Vio plasmid at increasing aTc concentrations 442 the addition of sponge sites reduces the occupancy of bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 8

Competitive binding FnCas12a S 465 d 1 DISCUSSION S A d 2 Sponge sites B

Sd CRISPRi 3 Tunable plasmid 1.2 Ptet copy number Nc ... crRNA Sd with decoy binding Sa Nc P sites Sd. 466 This work details the development of a system of tun- A 1.0 M sfGFP 467 able copy number plasmids and demonstrates their util- 0.8 150 468 ity in protein expression, biosynthetic pathway optimiza- Nc = 0 Nc = 30 0.6 Nc = 50 469 tion, and synthetic biology applications. Using a combi- 100 Nc = 170 0.4 Nc = 0 Nc = 270 Nc = 30 470 nation of high-throughput measurements and single-cell Nc = 50

50 Fluorescence Level 0.2 Nc = 170 471 experiments, we deduce relations between plasmid copy Nc = 270 Fluo. Intensity 0.0 472 number and host cells growth rate, protein expression, 0 [aTc] = 1ng/mL −4 −2 0 2 0 2 4 6 10 10 10 10 473 biosynthesis, and the performance of simple regulatory Time (h) aTc (ng/mL) C D 474 elements. This system could help resolve known issues 100 ON competitor site 475 currently plaguing the design and scaling up of synthetic 200 75 50 476 gene circuits such as leaky expression, narrow ranges of 25 active site 477 activation and repression control, and a large metabolic OFF Active site On/Off level 0 0

0 100 200 300 occupancy (%) 0 100 200 300 478 burden due to competition over a common pool of re- Sponge copy number Nc Sponge copy number Nc 479 sources [34–36].

480 Using a massively parallel directed mutagenesis ap- Fig. 6. Control of a CRISPRi Inverter with Competitor Bind- 481 proach, we find that the most straightforward and re- ing Sites. A) Top: Diagram showing how production of a dCas12a CRISPR RNA (crRNA) is controlled by an aTc- 482 liable way to change the plasmid copy number in the inducible promoter, and the dCas12a+crRNA duplex can ei- 483 ColE1 Origin is through mutations to the promoter con- ther bind to the active site Sa, which turns off production 484 trolling the priming RNA. Further, changes to the ma- of sfGFP, or to a decoy site Sd residing on a “sponge” plas- 485 chinery controlling inhibitory RNA levels are much less mid. Bottom: Fluorescence intensity vs time for CRISPR- 486 likely to give stable constructs given the fact that muta- dCas12a systems with variable numbers of sponge sites (Nc). 487 tions to the promoter controlling inhibition change the B) Induction curves for a CRISPRi system with sponge sites 488 sequence of the priming transcript can have drastic ef- on plasmid with varying copy numbers. C) On-Off levels for 489 fects on plasmid copy numbers through secondary struc- variable numbers of sponge sites. Both the ON and OFF fluo- 490 ture changes [37]. Overall, these results suggest that the rescence level of the inverter increase as the number of sponge 491 ColE1 replication machinery is more robust to changes in sites increases. The dynamic range decreases with increasing sponge copy number. D) Active site occupancy vs number 492 the priming RNA transcription initiation rate and much of sponge sites, as the number of sponge sites increases the 493 less robust to changes in the parameters detailing inhibi- active site becomes less occupied. The gray area represents 494 tion. solution to the fold-change model [14] for a dCas12a binding 495 Our multiplexed method of generating a plasmid copy energy equal to −13.25 kBT (top) and −12.25 kBT (bottom). 496 number library allows one to rapidly screen the effects 497 of plasmid copy number on a given system and could 498 be tailored to select for genetic devices that perform the 499 best under a given set of conditions. A key finding of this 443 the active site Sa and decreases the repression efficiency 500 facet of the work is that ColE1-type plasmids impose a 444 of the CRISPRi inverter. 501 linear metabolic burden on their hosts proportional to 445 We also observe a drastic reduction in the dynamic 502 the relative size of the plasmid, which may shed light on 446 range of circuits as the number of sponge sites is increased 503 the evolution of various regulatory elements in inhibitor- 447 (Fig. 6B), where cells with more than 150 sponge sites 504 dilution copy number control mechanisms. 448 have at least a 40% reduction in dynamic range compared 505 Our single cell measurements also show that plasmid 449 to those with no sponge sites. In contrast, the addition 506 loss and runaway replication are quite common phenom- 450 of 30 to 50 sponge sites only reduces the dynamic range 507 ena. These findings are in line with recent single cell 451 by approximately 6-8% (Fig. 6C). 508 measurements of plasmid copy numbers [6] and illustrate

452 While a large dynamic range is often desirable in ge- 509 that very careful attention should be given to the plas- 453 netic devices, a reduction in dynamic range could be used 510 mids used to house genetic circuits in order to ensure re- 454 to divert the flux in a metabolic pathway, ensure that the 511 liability. Since we observed an increasingly large fraction 455 concentration of some protein stays within a given range, 512 of non-growing cells as we increased the plasmid copy 456 or to reduce the toxicity of CRISPR-Cas proteins [33] due 513 number, our results further suggest that lower growth 457 to spurious off-target binding. Using a simple model of 514 rates population carrying a high-copy number plasmid 458 protein-DNA binding [14], we can determine that the oc- 515 may be explained by non-growing cells as opposed to a 459 cupancy of the active binding site Sa (Fig. 6D) in the 516 universal reduction in cellular growth rates. While this 460 OFF state is reduced from nearly 100 % with zero sponge 517 work is restricted to the ColE1 origin of replication, sim- 461 sites to 50 % when 270 sponge sites are present. Overall, 518 ilar principles can be used to tune the copy numbers of 462 we demonstrate that plasmid copy number can be used to 519 other inhibitor-dilution controlled plasmids. 463 accurately tune the level of transcription factor titration 520 The implications of plasmid copy number go beyond 464 in a simple genetic inverter. 521 its effects on gene expression and host cell fitness that bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 9

522 we’ve presented thus far. As vehicles of horizontal gene 576 The transformation efficiency was determined by plat- 523 transfer, the prevalence of which increases with plasmid 577 ing of serial dilutions of recovered cells. After trans- 524 copy number, plasmids play a key role in bacterial ecol- 578 formation liquid cultures were grown in Terrific Broth 525 ogy and evolution[38]. On the other hand, a consequence 579 (TB, VWR) until reaching an OD600 of 1.0, at which 526 of a high copy number is a reduced retention and fixation 580 point cultures were diluted 1:4 and regrown to OD600 527 of novel plasmid variants [39]. Together, these two issues 581 of 1.0, a procedure that was repeated 4 times. Excess 528 create a complicated image of the role of plasmid copy 582 culture from each time point was set aside for plasmid 529 number in bacterial diversity and evolution, it’s possible 583 isolation. Plasmids were isolated by miniprep (Zymo) 530 that the methods for controlling plasmid copy number 584 and a 243 bp region of the ori was amplified (pBR322- 531 presented here can be used help shed light on the role of 585 5prime-promoter1-rev 5’-TCTACACTCT TTCCCTA- 532 plasmid replication in bacterial evolution. 586 CAC GACGCTCTTC CGaTcTCAGA CCCCGTA- 533 The evolution of the fitness cost associated with 587 GAA AAGaTcAAAG GaTcTTC-3’, pBR322-3prime-for 534 plasmid-borne antimicrobial resistance genes is of cen- 588 5’-GACTGGAGTT CAGACGTGTG CTCTTCCGAT 535 tral importance in preventing the spread of drug-resistant 589 CTCGAGGTAT GTAGGCGGTG CTACA-3’) via PCR, 536 bacteria [40]. Our work presented here could provide 590 at which point universal primers and indices were at- 537 a clear path for investigating the role of plasmid copy 591 tached to the segment using an additional PCR step 538 number in the evolution of the cost of antimicrobial re- 592 (primers). Prepared libraries were then diluted to 50 pM 539 sistance. Further, the fitness cost of plasmid expression 593 and sequenced on the Illumina iSeq using overlapping 540 is highly species-dependent and dependent on the diver- 594 2 × 150 bp overlapping reads. 541 sity of a bacterial community [38] and recent findings 542 have shown that plasmids can quickly achieve fixation 543 in a population under non-selective conditions [41], fur- 595 Construction of inducible plasmids 544 ther understanding of plasmid replication mechanisms 545 and copy number effects might allow for a more nuanced 596 The pUC-pTet plasmid, whose copy number is in- 546 understanding of these varied properties of plasmids. 597 ducible in the presence of aTc was constructed using a 547 Together, these findings demonstrate that plasmid 598 PCR-based insertion as follows. Primers (pTet-tn10-for 548 copy number can have drastic effects on host cellular pro- 599 5’-AAATAACTCT aTcAATGATA GAGTGTCAAG 549 cesses and provides straightforward ways of investigating 600 AAGaTcCTTT GaTcTTTTCT ACGGGGTCTG A- 550 these effects in other systems. Our results underscore 601 3’, pTet-tn10-rev 5’-TACCACTCCC TaTcAGTGAT 551 the importance of tuning plasmid copy number as tool 602 AGAGATaTcT GCAAACAAAA AAACCACCGC 552 for the optimization of biosynthetic pathways, gene reg- 603 TACCAGC-3’) overlapping with the pUC19 origin of 553 ulatory circuits, and other synthetic biological systems. 604 replication and containing a pTet-inducible promoter 605 were used to linearize and amplify the pUC19 backbone. 606 The linearized plasmid was then circularized using a 554 MATERIALS AND METHODS 607 KLD reaction (NEB) and transformed into competent 608 cells via electroporation. During the recovery period 555 Next-generation sequencing library preparation 609 1x aTc was added to the recovery media to facilitate 610 replication of the plasmid.

556 The multiplexed libraries of plasmids was generated by 611 The pUC-pLac-RNAi plasmid, whose copy number 557 site-directed mutagenesis using the NEB Q5 High-fidelity 612 can be reduced through the addition of IPTG, was con- 558 polymerase. Primers (pBR322-promoter1-1024-rev 5’- 613 structed as follows. The inhbitor RNA (commonly called 559 GAKKKMAAGA AGaTcCTTTG aTcTTTTCTA 614 RNAI) from the pUC19 plasmid was isolated via PCR 560 CGGGGTCTG-3’,pBR322-promoter1-1024-for 5’- 615 (primers), in the process 15 bp overhangs were added 561 CTTTTTTTCT GCGCGTWWRM TGCTGCTNGC 616 to facilitate Gibson assembly. Another pUC19 plasmid 562 AAACAAAAAA ACCACCG-3, pBR322-promoter1- 617 was the linearized downstream of the pLac promoter 563 1024-rev 5’-GTTAGGCCAC CAKKKMAAGA ACTCT- 618 (primers) and the linearized plasmid together with the 564 GTAGC ACCGCCTACA TACC-3’, pBR322-promoter1- 619 inhibitory RNA insert were assembled via Gibson assem- ◦ 565 1024-for 5’-TACGGCTWWR MTAGAAGNAC AG- 620 bly at 50 C for 1 h and transformed into competent cells 566 TATTTGGT aTcTGCGCTC TGCTG-3’ ) containing 621 via electroporation. 567 degenerate nucleotides in the -35, -10, and +1 sites of 568 the promoters controlling transcription of both RNAs 569 were used to amplify the standard pUC19 plasmid 622 Copy number and growth rate measurement via 623 570 (NEB). After PCR amplification samples were digested next-generation sequencing reads 571 with DpnI (NEB) restriction enzyme to remove excess 572 pUC19 template. 624 To determine the relative plasmid copy number from 573 The PCR product was then circularized with electroli- 625 the abundance of next generation sequencing reads we 574 gase (NEB) and transformed with high efficiency (≥ 1.5 626 compare the fraction of the sequencing reads at several 575 million CFU/mL) into Endura Electrocompetent cells. 627 time points with that in the initial library. The ratio of bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 10

628 the two, with a small correction applied to account for 665 Absolute Copy Number Determination via digital 629 differences in the growth rates of cells harboring different 666 droplet PCR (ddPCR) 630 plasmids, is what we report as a relative plasmid copy

631 number. 667 To make absolute determinations of plasmid copy num- 632 We measure initial fraction of the reads in the ligation 668 bers we used digital droplet PCR. Colonies hosting the 633 product (fi) and the final fractions of the reads of the 669 appropriate constructs were picked off of plates after 634 plasmids harvested at different time points (ff,f2,f3,f4). 670 transformation and grown in 3 mL of TB or H me- −1 −1 635 Differences in the doubling time () can be measured from 671 dia (10 g L tryptone, 8 g L NaCl) to OD600 of 1.0. 636 differences in the fractions of reads from points separated 672 Cells were then serially diluted 8000-fold in nuclease free ◦ 637 in time. These relationships can be described in the equa- 673 ddH2O. Cells were lysed by heating at 95 C for 20 min. 638 tions below. First, to determine the doubling time from 674 After lysis 1 µL of the lysate was used in each digital PCR 639 adjacent time points: 675 reaction. Two digital PCR reaction were performed per 676 sample - one targeting the genomic dxs gene and an- δt( 1 − 1 ) fn = fn−12 τ τav 677 other targeting the beta lactamase, bla, gene on the plas- 678 mid (primers). This defines the plasmid copy number 679 as the number of plasmids per genome. After generat- 640 The above line states that the fraction at a given time 680 ing droplets on the QX200 droplet generator (Biorad), 641 point is equal to the fraction at the previous time point ◦ 681 reactions were thermal cycled at 95 C for 10 min, fol- 642 multiplied by the difference in the number of doublings ◦ ◦ 682 lowed by 35 cycles of 98 C for 30 s, 58 C for 30 s, and 643 from the bulk culture. Below we have rearranged the ◦ ◦ 683 72 C for 1 min, the a 1 min extension at 72 C. Sam- 644 equation to solve for the growth rate from the variables ◦ 684 ples were held at 4 C after the reaction was completed 645 we can either observe or control. 685 and before droplets were read. Droplets were read with 686 the QX-200 Droplet Reader (Biorad) and concentrations

fn 687 were determined using the Quantasoft analysis pro soft- Log2( f ) 1 1 n−1 + = 688 ware (Biorad). δt τav τ

646 Secondly, to determine the copy number from the 689 PCN Sponge Effects on CRISPRi Inverter 647 growth rate and the difference in sequencing counts we 648 express the fraction at the first time point as the product 690 CRISPR system with nuclease activity deactivated Cas 649 of the fraction in the ligation, the relative copy number, 691 protein(dCas) can be used to interfering transcription 650 and the number of excess doubling times: 692 regulation(CRISPRi) by design crRNA targeting at the 693 promoter for the gene of interest which acts a repressor t( 1 − 1 ) 694 of gene. However, dCas/crRNA complex can also bind to ff = fiC2 τ τav 695 other site that shares the same DNA sequence as the tar- 651 696 get but not directly regulating the gene(competitor site). 697 When the second scenario happens, it will have an impact fn −t( 1 − 1 ) 2 τ τav = C 698 on the repression level of the gene at given concentration fi 699 of dCas/crRNA, i.e., a sponge effect for the gene of inter- 700 est by consuming dCas/crRNA from the same pool. To 652 By inserting the above expression for the doubling 701 test the sponge effect at different number of competitor 653 time, we can have the relative copy number solely in 702 sites, a dual-plasmid system was used. The main plas- 654 terms of the fractions of each sequencing reads and the 703 mid(Kanamycin resistant) is a low copy(pSC101) pTet- 655 ratio of the time between time points and the time be- 704 inducible CRISPRi-dFnCas12a inverter, where the cr- 656 tween transformation and the first time point. 705 RNA transcribes by pTet target at a promoter that 706 drivers a green fluorescence protein(sfGFP). The second 707 plasmid is edited from the above mentioned pUC19-PCN t ff fn−1 − 708 library by inserting a single competitor via standard Site- 2 δt = C fi fn 709 Directed Mutagenesis(NEB, E0554). Two plasmids were 710 co-transformed into GL002 strain that has dFnCas12a 657 This analysis hinges on the assumption that upon 711 and TetR integrated in its genome with Mix & Go! E.coli 658 transformation each cell receives a single plasmid, that 712 Transformation Kit(T3001). N different PCN-plasmid 659 cells do not transfer plasmids between each other, and 713 were tested with the same main inverter plasmid. The 660 that the number of transformants of a given plasmid is 714 expression of sfGFP from dual plasmid system with dif- 661 proportional to the amount of that construct present in 715 ferent PCN was individually measured at a gradient of 662 the initial electroligation product. Further, this analy- 716 20 aTc concentrations on the BioTek (Synergy H1) plate 663 sis assumes that cells harboring different constructs have 717 reader in H medium. Three biological replicates were 664 identical lag times. 718 used for each system, i.e., three different single colony bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 11

719 were picked up and grown in 2 mL H media without aTc 763 Single cell microscopy 720 for approximately 18 hours. Then, 100 µL of overnight 721 cell culture was used and diluted into 2 mL fresh H media 764 The single cell data was collected within a month over 722 before transferring to a 96 well plate where each of the 765 8 subsequent experiments with varying aTc concentra- 723 middle of 10 x 6 wells contains 1 L of diluted cell cul- µ 766 tions. Each experiment was run for 1300 to 5000 min, 724 ture and 200 L fresh H media with corresponding aTc µ 767 corresponding to at least 20 to 80 generations. All exper- 725 concentration. 768 iments were started from the same stock bacterial colony, −1 769 initially grown to OD600 0.5 at 10 ng mL aTc and kept 770 refrigerated. In order to start the experiment, 3 µL of 771 the stock colony was disperse in 3 mL of H media with

726 Growth Rate Measurements 772 plasmid selecting antibiotic carbenicillin and aTc. The 773 cells were grown to OD600 0.5±0.1, concentrated by cen- 774 trifugation and inserted in plasma cleaned microfluidic 727 Growth rates of individual constructs were measured 775 chips. 728 on the BioTek (Synergy H1) plate reader. Overnight cul- 776 After entering the chip bacteria populated and grew 729 tures of cells grown to saturation were serially diluted 777 in 60 µm long, 7.5 µm wide and 1.05 µm thick chambers 730 by a factor of 100,000 into well-mixed 200 µL aliquots of 778 connected to a supply channel flowing fresh media pres- 731 media containing the appropriate antibiotics and inducer ◦ 779 surized to 4 psi [22]. On top of H media, carbenicillin 732 concentrations. Cells were grown at 37 C for 36 h while −1 780 and aTc, 0.1 g L of bovine serum albumin was added 733 to OD600 and fluorescence (if appropriate) was recorded 781 to the media to improve evacuation of cells from the sup- 734 every 3 min. Growth rates were calculated from the re- 782 ply channel [42]. 735 sultant time series data by averaging a window of ± 6 min 783 Single cell were resolved a by epi-fluorescence mi- 736 around the maximal growth rate, calculated as a discrete 784 croscopy of sf-GFP with a 100 x, 1.4 NA apochromat Le- 737 difference of OD600 readings between time points. 785 ica objective. Single cell detection and tracking enabled 786 measurement of cell dimensions, intensity and growth 787 rate overtime. Cells can be classified in three categories: 788 (1) non dividing dark cells (loss of antibiotic resistance 738 Gene Expression Measurements 789 due to plasmid loss), (2) healthy cells and (3) non grow- 790 ing bright cells (likely due to plasmid overload). Whereas

791 739 The expression of green fluorescent protein (sfGFP) in bulk pathological cells of type (1) or (3) would quickly 792 740 was used a measure of gene expression in this work. Flu- be overgrown by healthy cells, in chambers, we find that 793 741 orescence measurements were taken on cells prepared as they can get stuck at the back or even populate the whole 794 742 described above as they grew. Gene expression was cal- chamber, rendering the chamber unusable when the cells 795 743 culated as follows - the average fluorescence in a small eventually die. This is particularly true for non divid- 796 744 window surrounding the maximal growth rate was cal- ing dark cells at low aTc concentration. By using 7.5 µm 797 wide chambers, we allow many cells to grow side by side, 745 culated and divided by the integral of the OD600 in the 798 746 same window. This can be interpreted as a the fluores- which increases the chances of healthy cells to repopulate 799 747 cence (ie gene expression) per cell and should be indepen- the chambers. However, this is not enough to maintain 800 748 dent of the growth rate since we divide by the integral healthy chambers for more than a few hours at low aTc −1 801 concentration. For aTc concentrations of 1 ng mL and 749 of the OD600. Six replicates of each inducer concentra- 802 750 tion were taken and the average and standard deviation above a majority of chambers last for the whole experi- 803 751 of this measure of gene expression were reported. ment time of 5000 min. We base intensity measurements 804 on growing cells only (growth rate larger than 1/3 of 805 mean growth rate) to prevent over representation of non 806 growing cells that can reach up to 40% of the detected 807 cells (Fig. 3F). Selected cells are highlighted in supple- 752 Violacein Production Measurements 808 mentary movie.

809 Due to the rapid loss of chambers we are unable to 753 Cultures of single colonies hosting plasmids containing 810 measured a stabilized fluorescence level for aTc concen- −1 754 the violacein biosynthesis pathway were grown overnight 811 tration below 1 ng mL . However, we can still observe 755 in Terrific Broth with appropriate antibiotic. First, 812 that the aTc concentration promote plasmid replication. 756 200 µL of liquid culture was centrifuged and cells were 813 Indeed, the chamber survive longer at higher aTc con- 757 resuspended in 50 µL of water. Then, 450 µL of methanol 814 centration (plasmid loss is less frequent) and the average 758 was added to the resuspended cells and they were shaken 815 fluorescence increases with aTc. For aTc concentration −1 −1 759 for 1 h at room temperature. Cells were spun down and 816 of 1 ng mL to 100 ng mL time dependent fluorescence 760 200 µL of the supernatant was transferred to a 96-well 817 and growth rate curves (Figure S2) stabilize after about 761 plate where the absorbance at 650 nm was used to quan- 818 600 min (i.e. 12 generations) and are expected to accu- 762 tify violacein production. 819 rately represent PCN in exponential growth condition. bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 12

820 All experiments and cell cultures were performed at 822 ACKNOWLEDGMENTS ◦ 821 37 C.

823 We wish to thank David Specht for helpful discussions 824 and critical review of the manuscript. We also thank 825 members of the Lambert laboratory for providing feed- 826 back on the manuscript. This work was supported by 827 NIH funding under 1R35 GM133759 Maximizing Inves- 828 tigators’ Research Award (MIRA).

829 [1] A. P. Togna, M. L. Shuler, and D. B. Wil- 880 Biotechnology 33, 78 (2016). 830 son, Effects of Plasmid Copy Number and Run- 881 [12] A. Rozkov, C. A. Avignone-Rossa, P. F. Ertl, P. Jones, 831 away Plasmid Replication on Overproduction 882 R. D. O’Kennedy, J. J. Smith, J. W. Dale, and M. E. 832 and Excretion of -Lactamase from Escherichia 883 Bushell, Characterization of the metabolic burden on Es- 833 coli, Biotechnology Progress 9, 31 (1993), eprint: 884 cherichia coli DH1 cells imposed by the presence of a plas- 834 https://aiche.onlinelibrary.wiley.com/doi/pdf/10.1021/bp00019a005.885 mid containing a gene therapy sequence, Biotechnology 835 [2] K. Nordstr¨omand B. E. Uhlin, Runaway-replication plas- 886 and Bioengineering 88, 909 (2004). 836 mids as tools to produce large quantities of proteins from 887 [13] F. M. Weinert, R. C. Brewster, M. Rydenfelt, R. Phillips, 837 cloned genes in bacteria, Bio/Technology (Nature Pub- 888 and W. K. Kegel, Scaling of Gene Expression with 838 lishing Company) 10, 661 (1992). 889 Transcription-Factor Fugacity, Physical review letters 839 [3] M. Dwidar and Y. Yokobayashi, Riboswitch Signal Am- 890 113, 258101 (2014). 840 plification by Controlling Plasmid Copy Number, ACS 891 [14] R. C. Brewster, F. M. Weinert, H. G. Garcia, D. Song, 841 Synthetic Biology 8, 245 (2019), publisher: American 892 M. Rydenfelt, and R. Phillips, The Transcription Factor 842 Chemical Society. 893 Titration Effect Dictates Level of Gene Expression, Cell 843 [4] R. U. Sheth, S. S. Yim, F. L. Wu, and H. H. Wang, Multi- 894 156, 1312 (2014). 844 plex recording of cellular events over time on CRISPR bi- 895 [15] J. Paulsson, K. Nordstr¨om,and M. Ehrenberg, Require- 845 ological tape, Science 358, 1457 (2017), publisher: Amer- 896 ments for Rapid Plasmid ColE1 Copy Number Adjust- 846 ican Association for the Advancement of Science Section: 897 ments: A Mathematical Model of Inhibition Modes and 847 Report. 898 RNA Turnover Rates, Plasmid 39, 215 (1998). 848 [5] L. Potvin-Trottier, N. D. Lord, G. Vinnicombe, and 899 [16] G. del Solar, R. Giraldo, M. J. Ruiz-Echevarr´ıa,M. Es- 849 J. Paulsson, Synchronous long-term oscillations in a syn- 900 pinosa, and R. D´ıaz-Orejas,Replication and Control of 850 thetic gene circuit, Nature 538, 514 (2016). 901 Circular Bacterial Plasmids, Microbiology and Molecular 851 [6] B. Shao, J. Rammohan, D. A. Anderson, N. Alper- 902 Biology Reviews 62, 434 (1998). 852 ovich, D. Ross, and C. A. Voigt, Single-cell measurement 903 [17] J.-i. Tomizawa, Control of plasmid replication: Ini- 853 of plasmid copy number and promoter activity, Nature 904 tial interaction of RNA I and the primer transcript is 854 Communications 12, 1475 (2021). 905 reversible, Cell 40, 527 (1985). 855 [7] J. C. Diaz Ricci and M. E. Hern´andez, Plasmid Ef- 906 [18] J. Paulsson and M. Ehrenberg, Noise in a minimal regu- 856 fects on Escherichia coli Metabolism, Critical Reviews 907 latory network: plasmid copy number control, Quarterly 857 in Biotechnology 20, 79 (2000). 908 Reviews of Biophysics 34, 1 (2001), publisher: Cam- 858 [8] A. S. Karim, K. A. Curran, and H. S. Alper, 909 bridge University Press. 859 Characterization of plasmid burden and copy 910 [19] J. Paulsson, Multileveled selection on plasmid replica- 860 number in Saccharomyces cerevisiae for opti- 911 tion., Genetics 161, 1373 (2002). 861 mization of metabolic engineering applications, 912 [20] M. Plotka, M. Wozniak, and T. Kaczorowski, Quan- 862 FEMS Yeast Research 13, 107 (2013), eprint: 913 tification of Plasmid Copy Number with Single Colour 863 https://onlinelibrary.wiley.com/doi/pdf/10.1111/1567- 914 Droplet Digital PCR, PLoS ONE 12, 10.1371/jour- 864 1364.12016. 915 nal.pone.0169846 (2017). 865 [9] D. S.-W. Ow, P. M. Nissom, R. Philp, S. K.-W. Oh, 916 [21] P. Wang, L. Robert, J. Pelletier, W. L. Dang, F. Taddei, 866 and M. G.-S. Yap, Global transcriptional analysis of 917 A. Wright, and S. Jun, Robust Growth of Escherichia 867 metabolic burden due to plasmid maintenance in Es- 918 coli, Current Biology 20, 1099 (2010). 868 cherichia coli DH5 during batch fermentation, Enzyme 919 [22] G. Lambert and E. Kussell, Memory and Fitness Op- 869 and Microbial Technology 39, 391 (2006). 920 timization of Bacteria under Fluctuating Environments, 870 [10] W. E. Bentley, N. Mirjalili, D. C. Andersen, R. H. 921 PLoS Genet 10, e1004556 (2014). 871 Davis, and D. S. Kompala, Plasmid-encoded pro- 922 [23] D. A. Specht, Y. Xu, and G. Lambert, Massively parallel 872 tein: The principal factor in the “metabolic burden” 923 CRISPRi assays reveal concealed thermodynamic deter- 873 associated with recombinant bacteria, Biotechnol- 924 minants of dCas12a binding, Proceedings of the National 874 ogy and Bioengineering 35, 668 (1990), eprint: 925 Academy of Sciences 117, 11274 (2020). 875 https://onlinelibrary.wiley.com/doi/pdf/10.1002/bit.260350704.926 [24] H. C. Lichstein and V. F. Van De Sand, Violacein, an 876 [11] M. Pasini, A. Fern´andez-Castan´e, A. Jaramillo, 927 Antibiotic Pigment Produced by Chromobacterium Vio- 877 C. de Mas, G. Caminal, and P. Ferrer, Using promoter 928 laceum, The Journal of Infectious Diseases 76, 47 (1945), 878 libraries to reduce metabolic burden due to plasmid- 929 publisher: Oxford Academic. 879 encoded proteins in recombinant Escherichia coli, New 930 [25] M. Kholany, P. Tr´ebulle, M. Martins, S. P. Ventura, bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 13

931 J. Nicaud, and J. A. Coutinho, Extraction and purifica- 995 [42] D. Yang, A. D. Jennings, E. Borrego, S. T. Retterer, and 932 tion of violacein from Yarrowia lipolytica cells using aque- 996 J. M¨annik,Analysis of factors limiting bacterial growth 933 ous solutions of surfactants, Journal of Chemical Tech- 997 in pdms mother machine devices, Frontiers in Microbiol- 934 nology & Biotechnology , jctb.6297 (2019). 998 ogy 9, 871 (2018). 935 [26] D. G. Gibson, L. Young, R.-Y. Chuang, J. C. Venter, 936 C. A. Hutchison, and H. O. Smith, Enzymatic assem- 937 bly of DNA molecules up to several hundred kilobases, 938 Nature Methods 6, 343 (2009). 939 [27] M. R. Atkinson, M. A. Savageau, J. T. Myers, and A. J. 940 Ninfa, Development of Genetic Circuitry Exhibiting Tog- 941 gle Switch or Oscillatory Behavior in Escherichia coli, 942 Cell 113, 597 (2003). 943 [28] Y. Mileyko, R. I. Joh, and J. S. Weitz, Small-scale copy 944 number variation and large-scale changes in gene expres- 945 sion, Proceedings of the National Academy of Sciences 946 105, 16659 (2008). 947 [29] T. Wang, N. Tague, S. A. Whelan, and M. J. Dunlop, 948 Programmable gene regulation for metabolic engineer- 949 ing using decoy transcription factor binding sites, Nucleic 950 Acids Research 49, 1163 (2021). 951 [30] B. Wang, F. Guo, S.-H. Dong, and H. Zhao, Activation of 952 silent biosynthetic gene clusters using transcription factor 953 decoys, Nature Chemical Biology 15, 111 (2019). 954 [31] M. Z. Ali and R. C. Brewster, Controlling gene expression 955 timing through gene regulatory architecture, bioRxiv , 956 2021.04.09.439163 (2021). 957 [32] X. Wan, F. Pinto, L. Yu, and B. Wang, Synthetic protein- 958 binding DNA sponge as a tool to tune gene expression 959 and mitigate protein toxicity, Nature Communications 960 11, 5961 (2020). 961 [33] S. Zhang and C. A. Voigt, Engineered dCas9 with re- 962 duced toxicity in bacteria: implications for genetic circuit 963 design, Nucleic Acids Research 46, 11115 (2018). 964 [34] C. Tan, P. Marguet, and L. You, Emergent bistability 965 by a growth-modulating positive feedback circuit, Nature 966 Chemical Biology 5, 842 (2009). 967 [35] Y. Qian, H.-H. Huang, J. I. Jim´enez,and D. Del Vecchio, 968 Resource Competition Shapes the Response of Genetic 969 Circuits, ACS Synthetic Biology 6, 1263 (2017). 970 [36] F. Ceroni, R. Algar, G.-B. Stan, and T. Ellis, Quantifying 971 cellular capacity identifies gene expression designs with 972 reduced burden, Nature Methods 12, 415 (2015). 973 [37] J. Tomizawa and T. Itoh, Plasmid ColE1 incompatibility 974 determined by interaction of RNA I with primer tran- 975 script., Proceedings of the National Academy of Sciences 976 78, 6096 (1981). 977 [38] J. Rodr´ıguez-Beltr´an,J. DelaFuente, R. Le´on-Sampedro, 978 R. C. MacLean, and San Mill´an,Beyond horizontal gene 979 transfer: the role of plasmids in bacterial evolution, Na- 980 ture Reviews Microbiology 19, 347 (2021). 981 [39] J. Ilhan, A. Kupczok, C. Woehle, T. Wein, N. F. H¨ulter, 982 P. Rosenstiel, G. Landan, I. Mizrahi, and T. Dagan, Seg- 983 regational Drift and the Interplay between Plasmid Copy 984 Number and Evolvability, Molecular Biology and Evolu- 985 tion 36, 472 (2019), publisher: Oxford Academic. 986 [40] T. Vogwill and R. C. MacLean, The genetic 987 basis of the fitness costs of antimicrobial re- 988 sistance: a meta-analysis approach, Evolu- 989 tionary Applications 8, 284 (2015), eprint: 990 https://onlinelibrary.wiley.com/doi/pdf/10.1111/eva.12202. 991 [41] T. Wein, N. F. H¨ulter,I. Mizrahi, and T. Dagan, Emer- 992 gence of plasmid stability under non-selective conditions 993 maintains antibiotic resistance, Nature Communications 994 10, 2595 (2019). bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 1

1 Supplementary Materials 44 in Fig. 3B and 3C of the main text, as well as supple- −1 45 mentary figure 4, for aTc conditions of 1 ng mL and 46 above. −1 2 MOTHER MACHINE 47 For aTc concentration of 0.1 ng mL and below, plas- 48 mid loss is too frequent, we are not able to measure a 49 stabilized average cell intensity. We use all time points 3 Data selection 50 available for those concentrations in other plots across 51 the manuscript. 4 In order to discard pathological cells as well as detec- 5 tion errors the following threshold are used to select data.

52 Growth rate 6 1. 0.7 µm

−1 −1 7 2. 0.005 min < growth rate <0.04 min 53 Growth rate measurement particularly lack of preci- 54 sion for the low aTc conditions. Considering aTc con- −1 8 3. 1.7 µm 122 A.U. 57 dition of 100 ng mL .

10 The intensity threshold is set just a few unit above the 11 average background (110-120). The most restrictive con- 58 PCN distribution 12 ditions is the one on growth rate. It is used to set apart 13 non growing cells as well as cells that can not be tracked 59 In supplementary figure S5, we plot fold change his- 14 overtime due to detection errors. Supplementary figure 60 tograms and corresponding gamma distributions: 15 S1 shows histogram of cell properties before and after se- −1 16 lection for one of the worst cases at 0.01 ng mL . The 1 a−1 − x p(x) = x e b , (1) 17 peak at 0 correspond to cells without measured growth baΓ(a) 18 rate. Detection and tracking error may lead to cell

19 2 tracks which length decreases overtime leading to neg- k1 hxi 61 with Γ the gamma function, a = = 2 the mean 20 ative growth rate. γ2 σ 62 number of burst per cycle (Friedman PRL 2006) and b = 2 k2 σ 63 = the mean burst size. γ1 hxi 21 Cell morphology

64 MASSIVELY PARALLELIZED ASSAY 22 In supplementary figure S2, we compare histograms of 23 cells morphology for all aTc conditions. We observe that −1 −1 65 No Relationship Between Priming Promoter 24 the cells are longer at 10 ng mL aTc than 0.01 ng mL . 66 Strength and Plasmid Copy Number 25 However, the volume - surface area ratio is conserved.

67 Given that we saw a clear relation between promoter

68 26 Stabilization of PCN de-repression and plasmid copy number (Fig. 1D) and 69 that high-copy number plasmids had similar origins of 70 replication, we were surprised that we could not find a 27 In S3 we plot cell properties as a function of time for 71 clear relation between the predicted strength [14] of the 28 each aTc concentration. 72 promoter controlling RNA-p and the plasmid copy num- 29 The number of detected cells quickly increases when 73 ber (Fig. S6). 30 cells populate the chambers. Then, the total number of 31 detected cells decreases for all aTc conditions. This is 32 largely due to chambers being filled by pathological cells 33 that upon death block their chamber. Pathological cells 34 mostly suffer from plasmid loss at low aTc while plasmid 35 overload dominates at high aTc. We observe that plas- 36 mid loss is more prone to block chambers. Indeed, the 37 more aTc, i.e. the more PCN, the longer lived are the 38 chambers. −1 39 For aTc concentration of 1 ng mL and above, we 40 observe a stabilization of the average cell intensity af- 41 ter about 600 minutes (about 12 generations). We also 42 observe a stabilization in growth rate around this time. 43 Only the data points past this 600 min threshold are used bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 2

TABLE I. Table S1: Promoter Sequences, Next-generation sequencing counts, Relative Growth Rates, Predicted Plasmid Copy Numbers, and Predicted Promoter Strength for priming RNA variants used in this work

TABLE II. Table S2: Promoter Sequences, Next-generation sequencing counts, Relative Growth Rates, Predicted Plasmid Copy Numbers, and Predicted Promoter Strength for inhibitory RNA variants used in this work

TABLE III. Table S3: Sequencing Counts at each time point for priming RNA variants

120000 100000 after 120000 175000 100000 before 100000 150000 80000

80000 125000 80000 60000 100000 60000 60000 75000 40000 40000 40000 50000 20000 20000 20000 25000

0 0 0 0 0 1000 2000 3000 4000 0 5 10 15 20 0 500 1000 1500 2000 2500 3000 0.00 0.02 0.04 0.06 time length ( m) intensity growth rate (1/min)

Fig. S1. Count of cells as a function of time, length, fluorescence intensity, and growth rate before and after selection for one of the most difficult condition at 0.01 ng mL−1 aTc.

4 0.8 0.4 3 0.6

0.2 2 0.4 1 0.2

0.0 0 0.0 0.0 2.5 5.0 7.5 10.0 0.0 0.5 1.0 1.5 2.0 0 2 4 6 8 length ( m) diameter ( m) volume, V ( m3)

20 0.0 ng/mL 0.15 15 0.01 ng/mL 0.10 10 0.1 ng/mL

0.05 5 0.3 ng/mL 1.0 ng/mL 0.00 0 0 10 20 30 0.1 0.2 0.3 0.4 3.0 ng/mL surface area, SA ( m2) V/SA ( m) 10.0 ng/mL 100.0 ng/mL

Fig. S2. Increase of cell length with aTc concentration, i.e. plasmid copy number, but conserved volume/surface area ratio.

2500 7.5 20000 0.024 0.0 7.0 0.01 2000 6.5 0.022 0.1 ) 15000

m 6.0 0.020 1500 0.3 ( 5.5 0.018 1.0 h 10000 t 5.0 1000 g 3.0 0.016 n intensity

e 4.5 10.0 l 5000 detected cells 500 0.014 100.0 4.0 growth rate (1/min) 3.5 0.012 0 0

0 1000 2000 3000 4000 0 1000 2000 3000 4000 0 1000 2000 3000 4000 0 1000 2000 3000 4000 time (min) time (min) time (min) time (min)

Fig. S3. Time dependent measurement of the number of detected cells, their averaged length, fluorescence and growth rate. bioRxiv preprint doi: https://doi.org/10.1101/2021.07.13.451660; this version posted July 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-ND 4.0 International license. 3

0.04

0.03

0.02

growth rate (1/min) 0.01

0.0 0.01 0.1 0.3 1.0 3.0 10.0 100.0 aTc (ng/mL)

Fig. S4. Violin plots illustrating the distribution of growth rates for all aTc concentrations after selection (c.f. supplementary figure 1).

model fit (std+mean) model fit (std+mean)

5 Friedman_a=0.8_b=0.6 Friedman_a=1.8_b=0.5 1.0 0.8 3.0 4 0.6 3

0.4 density 2 density

1 0.2

0 0.0 0 1 2 3 4 0 1 2 3 4 5 fold change fold change

model fit (std+mean) model fit (std+mean) 1.0 Friedman_a=5.2_b=0.2 1.0 Friedman_a=6.4_b=0.2 0.8 10.0 100.0 0.8

0.6 0.6

density 0.4 density 0.4

0.2 0.2

0.0 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 fold change fold change

Fig. S5. Fit of fold change histograms for stabilized movies at 1, 3, 10 and 100 ng mL−1 aTc.

103

102

101

100 Plasmid Copy Number

10−1 −4 −3 −2 −1 0 1 2 3 Predicted Promoter Strength (kT)

Fig. S6. Priming Promoter Strength vs Plasmid Copy Number