Single-cell mass distributions reveal simple rules for achieving steady-state growth

Optical density is a common method for measuring exponential growth in bacterial batch cultures. However, there is a misconception that such exponential growth is equivalent to steady-state growth, which is a distinct physiological state that improves experimental reproducibility. Determining precisely when steady-state growth occurs is technically challenging and is aided by paired single-cell and population-level measurements. Using microfluidic mass sensors and optical density, we explore when in typical laboratory batch cultures steady-state growth occurs. We show that cell mass increases by an order of magnitude within a few hours of dilution into fresh medium and that steady-state growth is only achieved when cultures are inoculated with high dilutions from overnight stationary phase cultures. At high dilutions, Escherichia coli and Vibrio cyclitrophicus grown in different rich media achieve steady-state growth approximately 4 total biomass doublings after inoculation. We can decompose these dynamics into 3 doublings of average cell mass and 1 doubling of cell number for both species. We also show that batch cultures in rich media depart steady-state growth early in their growth curves at low cell and biomass concentrations. Achieving and maintaining steady-state growth in batch culture is a delicate balancing act, and we provide general guidance for commonly used rich media. Quantifying single-cell mass outside of steady-state growth is an important first step towards understanding how microbes grow in their natural context, where fluctuations pervade at the scale of individual cells. Importance Microbiologists have watched clear liquid turn cloudy for over 100 years. While the cloudiness of a culture is proportional to its total biomass, growth rates using such optical density measurements are challenging to interpret when cells change size. Many bacteria adjust their size at different steady-state growth rates, but also when shifting between starvation and growth. Optical density cannot disentangle how mass is distributed among cells of different sizes, and directly measuring how mass is distributed among cells has been a major challenge. Here we use single-cell mass measurements to demonstrate that a population of cells in batch culture achieves a stable mass distribution for only a short period of time. Achieving steady-state growth in rich medium requires low initial biomass concentrations and enough time for the coordination of individual cell and population growth. Steady-state growth is important for reliable cell mass distributions in a culture and we discuss how mass variation outside of steady-state can impact physiology, ecology, and evolution experiments.

cultures steady-state growth occurs. We show that cell mass increases by an order of 25 magnitude within a few hours of dilution into fresh medium and that steady-state growth is 26 only achieved when cultures are inoculated with high dilutions from overnight stationary 27 phase cultures. At high dilutions, Escherichia coli and Vibrio cyclitrophicus grown in different 28 rich media achieve steady-state growth approximately 4 total biomass doublings after 29 inoculation. We can decompose these dynamics into 3 doublings of average cell mass and 1 30 doubling of cell number for both species. We also show that batch cultures in rich media 31 depart steady-state growth early in their growth curves at low cell and biomass 32 concentrations. Achieving and maintaining steady-state growth in batch culture is a delicate 33 balancing act, and we provide general guidance for commonly used rich media. Quantifying Microbiologists have watched clear liquid turn cloudy for over 100 years. While the 39 cloudiness of a culture is proportional to its total biomass, growth rates using such optical 40 density measurements are challenging to interpret when cells change size. Many bacteria 41 adjust their size at different steady-state growth rates, but also when shifting between 42 starvation and growth. Optical density cannot disentangle how mass is distributed among 43 cells of different sizes, and directly measuring how mass is distributed among cells has been 44 a major challenge. Here we use single-cell mass measurements to demonstrate that a 45 population of cells in batch culture achieves a stable mass distribution for only a short 46 period of time. Achieving steady-state growth in rich medium requires low initial biomass 47 allows us to examine if there are general rules for achieving and departing from steady-state 96 growth among diverse bacteria in commonly used cultivation conditions. 97

98
The ability of a culture to achieve steady-state growth depends on the initial OD, that is 99 mass concentration, in the culture. We performed dilutions of 1:100, 1:1,000, or 1:10,000 100 from overnight stationary phase cultures into fresh medium to vary the initial OD, then 101 measured population dynamics over about 5 hours. The mean cell mass immediately 102 increased in all cultures with no apparent lag time upon encountering fresh medium ( Figure  103 1 a-f, Figure 2   The amount of mass gained by cells is enormous as they transition into steady-state growth 132 following starvation. While the mean buoyant mass change we observed could be 133 influenced by changes to the composition of cellular dry mass via altered dry density, this 134 effect is relatively small for E. coli in similar conditions(17). Therefore, E. coli buoyant mass 135 can be converted to dry mass using a multiplicative conversion factor between 3.1 (in 136 stationary phase) to 3.6 (near steady-state)(17). E. coli alters its dry mass more in the first 3 137 hours of our experiments (164fg-1,512fg), than it does across 20 different growth media 138 supporting steady-state growth rates between 0.31-1.72 per hour (240fg-1,180fg)(16). 139 Experimentalists comparing mass in multiple conditions (e.g., different media) must account 140 for this within condition variation by establishing steady-state growth. The mass range 141 clonal isolates can realize is vast, but cell mass is systematically responsive to the transient 142 and long-term(16) nutrient conditions bacteria experience. 143

144
The exit from steady-state occurs very early in the growth curve at low OD and cell 145 concentration. This has been reported previously for E. coli in LB(4, 5) and we observe 146 steady-state exit to occur at an even lower OD and cell concentration ( Microbiologists have known for a century that cell properties in liquid batch cultures are 178 dynamic(6), but exactly how much cells alter their mass over time is challenging to measure 179 directly. We believe this information gap, along with misconceptions about OD have led to 180 the common misunderstanding that an exponential OD increase is equivalent to steady-181 state growth. Here we demonstrate that cell mass changes rapidly and substantially in batch 182 culture, so care must be taken to achieve steady-state growth. While the exact amount of 183 mass added upon dilution into fresh medium will depend on the specific strain and medium 184 combination, it is a very large increase for these strains in commonly used rich medium. 185

186
We recommend that microbiologists intending to work with steady-state populations in 187 batch culture on rich media perform the following steps: 188 1) inoculate with a minimum of a 10,000-fold dilution from an overnight parent culture for 189 E. coli-like cell yields (~10 9 cells/ml). 25°C shaking at 250 RPM. The following cultivation protocol was followed for each 232 experiment. For each species, two biological replicate cultures were inoculated on day 1 233 from the same -80°C freezer stock into separate recovery cultures (5ml medium in 13ml test 234 tube) and allowed to grow overnight. We define biological replicate as an independent 235 culture originating from the same clonal freezer stock, but which was inoculated into a 236 separate recovery culture and maintained separately for all further transfers. On the 237 morning of day 2 the recovery cultures had all achieved stationary phase and high optical density, so were inoculated with a 1:100 dilution into pre-cultures (50µl into 5ml medium in 239 13ml test tube) and allowed to grow for an additional day. This ensured the cultures had 240 several generations of growth to recover from cryopreservation and were also in stationary 241 phase in the same medium used for experimentation for about 18 hours. Experimental 242 cultures were inoculated on day 3 using different dilutions (either 1:100, 1:1,000, or

Optical density measurements 254
Optical density at 600nm (OD600) was measured on a Genesys40 spectrophotometer 255 (Thermo Fisher) relative to uninoculated growth medium blanks in either cuvettes or test 256 tubes where appropriate. The accuracy limits of the spectrophotometer (0.01-0.79 OD600) 257 were determined with a dilution series of a stationary phase culture of V. cyclitrophicus 258 1G07 which was washed and re-suspended in a buffer that did not contain the 259 macronutrients necessary for growth. OD600 of cultures was measured in the side-arm of the 260 cultivation flask until approaching an OD600 value around 0.5, after which an additional 261 100µl of sample was destructively removed from the flask to dilute and measure the true 262 value withing the accuracy limit of the spectrophotometer. Briefly, these samples were 263 diluted 1:10 with fresh medium (0.1ml sample into 0.9ml medium) in a cuvette and 264 immediately measured in the same spectrophotometer with a cuvette adapter. Diluted 265 sample OD600 values were then multiplied by the dilution factor. 266 Batch cultures were prepared with medium that was filtered with 0.1µm pore size PES filters 276 immediately prior to use to ensure it was free of any small particles that could interfere with 277 cell measurements on the SMR cantilever. Negative controls of uninoculated medium were 278 run on the SMR prior to inoculation to ensure the particle background of each medium 279 batch was adequately low (concentrations of less than 1x10 5 background particles/ml, 280 comparable to ultrapure water controls). sensor simultaneously and only one particle's mass can be accurately measured, the 304 software registers 1 particle for mass purposes (1 measured particle) but 2 particles for 305 concentration purposes (2 detected particles). Therefore, the discrepancy in sample size for 306 different figure panels measured from the same sample relates to the difference between 307 measured and detected particles. 308