Next Article in Journal
Phosphate Versus Nitrogen Limitation: A Reactor-Scale Process Comparison for Single-Cell Oil Production in Oleaginous Yeasts
Previous Article in Journal
Effect of Aeration Process on Lignocellulosic Degradation, Humification and Carbohydrate-Active Enzyme (CAZymes) Genes in Aerobic Composting
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Fermentation
Volume 12
Issue 4
10.3390/fermentation12040171
fermentation-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Day in the Life of a Sourdough Leaven from Feeding to Maturity

by
Louis Levinger
1,2,*,
Monisha Sherpa
1,
Julia Gelman
3,†,
Mariapia Dibonaventura
1 and
Rabindra Mandal
4,5
1
Department of Biology, York College of The City University of New York, 94-20 Guy R. Brewer Blvd., Jamaica, NY 11451, USA
2
Doctoral Program in Biochemistry, Graduate Center of The City University of New York, 365 Fifth Avenue, New York, NY 10016, USA
3
Queens High School for the Sciences at York College, 94-50 159 Street, Queens, NY 11433, USA
4
Department of Biology, Hunter College of The City University of New York, 695 Park Avenue, New York, NY 10065, USA
5
The Doctoral Programs in Biology and Biochemistry, Graduate Center of The City University of New York, 365 Fifth Avenue, New York, NY 10016, USA
*
Author to whom correspondence should be addressed.
†
Julia Gelman is a student at Queens High School for the Sciences at York College and participated in the Urban Barcoding Research Program (UBRP) from 2024–2025.
Fermentation 2026, 12(4), 171; https://doi.org/10.3390/fermentation12040171
Submission received: 27 February 2026 / Revised: 15 March 2026 / Accepted: 18 March 2026 / Published: 24 March 2026
(This article belongs to the Section Fermentation for Food and Beverages)
Browse Figures
Review Reports Versions Notes

Abstract

Fermentation is a type of biological process conducted domestically or commercially to preserve foods and beverages, produce alcohol, add nutritional value and improve aroma and flavor. The natural fermentation of flour in water to obtain a leaven for baking, lately scrutinized in the laboratory with the application of metagenomic methods, has been ubiquitous since the dawn of civilization. Commercially, single culture or defined mixtures of microorganisms are used for their predictability, but regularly fed two-domain microorganism cultures are favored in less industrialized and domestic operations. Fungi principally produce the carbon dioxide responsible for leavening. The bacteria produce acid in the bread commonly known as sourdough for its aroma and flavor. A leaven made by fermentation using flour and water can be stored while it is dormant. We studied a mature culture that is fed twenty-fold with water and flour by incubating it for 24 h, sampling it regularly for pH measurements, and plating it. The colonies were suspended for micrography and DNA extraction for PCR and Sanger sequencing. The metagenomic DNAs were analyzed for bacterial and fungal composition. The proportions of the plant and microbial DNA endogenous to the flour decline rapidly, and the predominant bacteria and fungi in mature leaven propagate, without overlap between the respective microbiomes.

1. Introduction

Present-day commercial bread-baking is dominated by monoculture fermentation using the bakers’ yeast Saccharomyces cerevisiae. However, since prehistoric times, fermentative microbiomes have consisted of mixtures of fungi and bacteria worldwide [1,2], principally lactic acid bacteria (LAB). The benefits of mixing two domain fermentation microbiomes include an improved aroma, flavor, crust and crumb, a longer shelf life, and a lower glycemic index. The souring, due to the production of lactic acid by LAB [3,4], leading to an endpoint pH of 4 or slightly below, suppresses the spoilage microorganisms. Fungi produce carbon dioxide that is responsible for leavening, as well as ethanol and other volatile organics.
The stable coexistence of fungi and bacteria raises questions about their mutual benefits. Though experimentally unaddressed here, investigators using a systems-based design [5] posited that the LAB consume the fixed carbon in the form of disaccharides and more complex carbohydrates more capably than fungi, providing the fungi with simpler forms of fixed carbon via carbon overflow. Conversely, fungi were suggested to consume fixed nitrogen from the polymeric macromolecules, such as proteins and nucleic acids, more capably than bacteria, thereby supplying the LAB with simpler forms of fixed nitrogen, such as amino acids, via nitrogen overflow. The complete analysis would include a metabolomic consideration of how each member of the mixed microbial community degrades and consumes plant material, mediates overflow, and takes up the simpler forms of fixed carbon and nitrogen.
The plant plastid DNAs, detected in overwhelming proportions alongside the bacteria in the 16S flour metagenome, substantiate the endosymbiont hypothesis for the prokaryotic origin of mitochondria and chloroplasts in the eukaryotic progenote [6]. Fortuitously, 16S reads can be used to study the decreasing proportion of plant plastid DNAs, which were initially close to 100%, as a proxy for the consumption of flour as food by the fermentative microbiome. The ~500,000 Illumina paired-end reads also prove sufficient to analyze the endogenous bacterial microbiome in flour after the removal of plant plastid sequences from the datasets.
Grain berries and flour have similar endogenous microbiomes [2,3,7], as evidenced in our unpublished observations. The microorganisms accumulate coincidentally and remain dormant on the grain surfaces and in crevices. Grain berries remain viable, and flour retains its nutritional value when stored in a cool and dry place, becoming susceptible to spoilage only if stored improperly.
A mature starter can be stored indefinitely. Both wet and gently dehydrated starters are widely available. The constituent microorganisms go dormant when a starter culture saturates, is left unfed, and is stored in a refrigerator. A dormant starter can be rejuvenated by feeding it regularly with a fresh water–flour mixture.
A periodic two-fold feeding (back-slopping) is widely employed for leaven development, maintenance and expansion, in which half the culture is discarded and the retained culture is fed with an equal mass of a 50% (w/v) suspension of flour in water [1,2]. Alternatively, Robertson describes a twenty-fold feeding for leaven preparation [8]. For example, 1 g of a 50% (w/v) starter is suspended with added water to a mass of 10 g, and 10 g of dry flour mixture is mixed in. The present study, following Robertson’s feeding technique, enables a twenty-fold increase in cell densities, including an exponential growth phase, thus filling a gap in the understanding of the short-term natural history of leaven maturation. The saturation is observed after about 12 h at room temperature. The cycle begins with a high proportion of plant plastids and a low density of microorganisms endogenous to the flour, followed by a precipitous decrease in the proportion of plant material and the microbiome that is associated with the flour, corresponding with an increase in the proportion of the mature starter microbiome. The endogenous microorganisms that are present in the flour microbiome are practically absent from the 12 h culture, and conversely, the predominant microorganisms in the 12 h culture are practically absent from the flour.

2. Methods

Preparation and maintenance of a sourdough leaven or starter. A flour mixture was prepared from equal masses of all-purpose unbleached white (Heckers) and medium organic rye (King Arthur) flours. The flour mixture was initially hand mixed with an equal volume of water (e.g., 10 g flour mix, 10 g filtered water) and incubated either at home or in the laboratory until it became bubbly and started to rise. The glassware and mixing utensils at home were clean but non-sterile.
In contrast to the conventional two-fold feedings known as back-slopping [1,2], twenty-fold feedings were performed throughout the project [8], allowing for several generations of log phase growth. For example, 1 g of grown-out culture was retained, suspended in filtered water to a mass of 10 g, and 10 g of the white/rye flour mix was added and mixed until uniformity using a silicone rubber spatula.
To minimize the introduction of microorganisms from the exogenous environment, a starter was prepared in the laboratory (unpublished) using sterile tubes, sterile water and a sterile spatula for mixing. The cultures were fed with a frequency based on the rise observed, and by the fourth day, a regular daily rise was observed. Within a week, these cultures could be used to leaven bread and were subsequently maintained at home.
The laboratory preparation of leaven was initially conducted with two samples, one prepared from crushed grain and the other from commercial flour. Both of the samples could be used to leaven bread, but the crushed grain leaven displayed less of a rise, and its culture was discontinued.
A dormant culture of leaven that was prepared with flour in the laboratory was stored in the home refrigerator and fed and expanded at two-to-three-week intervals, which are required for bread making. Two twenty-fold feedings were performed at 12 or 24 h intervals to rejuvenate and expand the leaven. The last feeding was made with 10 g of leaven mixed with water to make 100 g following which 100 g of white/rye flour mixture was added to create a total of 200 g. Upon maturity, this was used in the Tartine recipe [8]; the mature culture that remained from the previous feeding was retained in the refrigerator.
Feeding and sampling over a 24 h time course. The fresh mature flour leaven was brought to the research lab, and 2 g was placed in a sterile 50 mL Falcon tube. Sterile deionized water was added to the 20 g and mixed with a sterile spatula. To create a total of 40 g, 20 g of a white/rye flour mixture was added and mixed to uniformity. Approximately 1 g of this mixture was dispensed into each of the 21 sterile 15 mL Falcon tubes, which is sufficient for seven triplicate samples, and was incubated at 27 °C. The triplicate samples of flour mixture (F) and the mature culture used for feeding (M) were also included. As presented, the samples are F, 0 h, 2 h, 4 h, 6 h, 8 h, 12 h, 24 h, and M.
In detail, ~250 mg samples were taken from the triplicate tubes at each time point for metagenomic DNA extraction and were stored frozen at −20 °C. The metagenomic DNAs were prepared after the collection of all the samples using the Qiagen Power Soil kit. The DNA concentrations, measured by Nanodrop, were typically 200 ng/μL in 70 μL.
Additionally, ~100 mg of each sample was weighed and diluted 10-fold w/v with sterile saline for the dilution series and plating. The fungi were grown on a potato dextrose agar (PDA) containing 200 μg/mL of chloramphenicol, incubated at 27 °C, and the bacteria were grown on a tryptic soy agar (TSA) containing cycloheximide at 200 μg/mL [9], incubated at 37 °C. Comparable results were obtained using an MRS agar (optimized for lactic acid bacteria). For flour suspensions, 100 μL of the 10−1, 10−2, and 10−3 dilutions were plated for both the fungi and the bacteria. At time = 0, for 2 h and 4 h samples, the 10−3 and 10−4 dilutions were plated for the fungi, and the 10−5 and 10−6 dilutions were plated for the bacteria. For the later time points and the mature sample, the 10−4 and 10−5 dilutions were plated for the fungi, and the 10−6 and 10−7 dilutions were plated for the bacteria. The plates were examined daily or after incubation over the weekend. Even the slowest growing colonies could be detected after three days of incubation.
Sterile water was added to the remainder of each sample to make 5 mL so that pH measurement could take place using a pH meter.
Colonies were used for wet mount microscopy and DNA extraction for PCR and Sanger sequencing. The colonies were sampled with a sterile micropipette tip and resuspended in 20 μL sterile saline for wet mount microscopy. The saline suspensions were loaded on a hemocytometer slide to provide scale markings. The best micrographs, taken at 400× for the fungi and 1000× with oil immersion for the bacteria, were obtained from colonies with actively dividing cells, after 1–3 days of incubation.
Colonies were also suspended in 50 μL 10% Chelex for DNA extraction (https://dnabarcoding101.org/, accessed on 14 November 2023). The Chelex suspensions were heated to 100 °C for 10 min and spun for 1 min at 1000× g in a microcentrifuge. Following this, 2 μL of the supernatant was added to the PCR tube containing a PCR bead (Cytiva) that was dissolved in a 23 μL primer mix using a 16S primer pair for the bacteria [10] or using an ITS primer pair for the fungi [11]. The primer sets include the M13F and R sequence tags at the 5′ ends for the Sanger sequencing. See Table 1 for primer sequences. The amplifications were performed with 35 cycles of 94 °C and 30’, 55 °C and 30’, and 72 °C and 2’, including an initial denaturation and a final extension step preceding and following the cycles. Following this, 2 μL of each PCR sample was electrophoresed on an agarose gel; those that amplified well were sent for Sanger sequencing using M13 forward and reverse primers (Azenta, South Plainfield, NJ, USA). The forward and reverse 16S bacterial sequences were assembled to ~1450 bp full-length amplicons. The amplicon sequences were assigned to genera using NCBI BLASTn (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch, accessed 14 November 2023) [12], and multiple sequence alignments were performed using MUSCLE [13]. Pantoea taxa, abundant in flour, possess indels close to the 5′ end of the Pantoea 16S amplicons, necessitating the design of a Pantoea-specific forward sequencing primer which anneals at nt 38–54 (Table 1).
Metabarcoding and data analysis. The metagenomic DNA samples were sent to SeqCenter (Pittsburgh, PA, USA) for microbial amplicon analysis, where they were split into two fractions for V3/V4 16S (bacterial) or ITS2 (fungal) amplicon library preparation using universal primers (Table 2). Briefly, for 16S rRNA gene analysis following the Qiime2 data analysis pipeline for microbiome data [14], adapters were removed [15] and sequences were denoised using QIIME2’s dada2 plugin [16]. The denoised sequences were used to construct the amplicon sequence variant (ASV) feature tables. The taxonomic classification of ASVs was performed using the Silva 138 99% full-length sequence database and the VSEARCH utility with QIIME2’s feature-classifier plugin. For ITS gene analysis, denoising and ASV feature table construction were performed as described above. ITS sequences were taxonomically assessed using the Unite 8 99% full-length sequence database and the VSEARCH utility within the QIIME2 plugin.
The high proportion of bacterial (especially Lactiplantibacillus plantarum) and fungal (especially Pichia) genera, which were inaccurately or inadequately classified to genus level by SeqCenter bioinformatic analysis, were manually corrected using NCBI BLASTn [12] with sequences obtained from the representative ASV Metadata.qzvs as queries (see Results/Interpretation). The taxonomic tables were exported to Excel and imported to MicrobiomeAnalyst v3.0 (www.microbiomeanalyst.ca) for visualization. The selected microbial proportion plots were prepared using the ggplot library on the R platform, generated using the tidyverse suite of R packages in RStudio. The microbial composition in the Excel format was imported into R using the readxl package and processed with dplyr and tidyr for data manipulation and transformation. For microbial community plots depicting specific bacterial and fungal taxa, relative abundance values were normalized to proportions summing to 1.0 for each sample. The additional plots, including the pH values and the fungal and bacterial cell density, were visualized without normalization. All of the proportion curves, including the bacterial and fungal taxa and cell density changes across the fermentation period, were generated using ggplot2. The logarithmic scaling was applied to the y-axis where appropriate to better visualize the rapid shifts in microbial proportions and the exponential growth in cell density.
Data access. The SeqCenter Fastq paired-end read files can be accessed at the NCBI SRA website PRJNA1426134. The link to the Github for data analysis and plotting in R is https://github.com/MonishaSherpa/Sourdough_Leaven_Metagenomics (accessed on 17 March 2026).
Student participation. Concurrent with the work presented here, more than 50 high school students and undergraduates participated in this and related projects from fall 2023 until summer 2025 (Table 3) and were drawn from the diverse racial, ethnic and cultural communities for which the Borough of Queens and the greater New York City region are known. Extramurally sponsored programs paid the student stipends, provided limited support for instructors, and funded the materials and sequencing services (see Acknowledgments). Culminating in the summer 2025 workshop, seven high school trainees worked with one instructor (LL) for the first three weeks and were joined by fifteen undergraduates with guidance from two instructors (LL and RM) and one recent undergraduate intern (MS) for the second half. Although participating students were not especially interested in the fermentative microbiomes at the outset, this project ideally integrates classical microbiological methods with bioinformatics.
The project content overlaps with the BS-level elective and the required Biology and Biotechnology major courses at York College/CUNY, including Biostatistics, Bioinformatics, Molecular Biology and Biotechnology, Microbiology, and Theory and Experimentation in Biotechnology. This project, which best fits the workshop context, helped a significant number of students by providing the opportunity for hands-on research combined with ample time for interpretation and discussion, as suggested by Table 3. In particular, high school students in structured programs are motivated and enthusiastic about participating in what for many of them is a first hands-on research opportunity.

3. Results and Interpretation

The lactic acid bacteria (LAB), mainly Lactiplantibacillus plantarum and Lactobacillus brevis, predominated in a leaven prepared from flour (see below). The greater diversity was initially observed in the fungal microbiomes of leavens prepared from crushed grain and flour. After one week, the crushed grain sample fungal microbiome was dominated by Mucor (unpublished results), a filamentous fungus prevalent in Taiwanese stinky tofu [17], while the flour leaven microbiome was composed of three abundant taxa: Alternaria, Blumeria and the family Cladosporiaceae. Saccharomyces was initially absent from both leavens. When imported home from the laboratory for leavening and baking, with nonsterile handling, Saccharomyces displaced the three abundant fungal taxa observed earlier in the flour leaven. Pichia, a yeast that has been widely used as a host for protein expression [18], later joined the fungal microbiome and gradually increased in proportion over Saccharomyces. The changes in microbiomes apparently originate from the exogenous home environment.
Leaven pH and cell density as a function of culture time. The flour leaven was brought back to the lab for 24 h time courses. The pH of ferments decreases with time due to the increase in density and the proportion of lactic acid bacteria (LAB) and, in some cases, acetic acid bacteria (AAB). Both bacterial and fungal microbiomes that are characteristic of the starting material (an equal mass mixture of unbleached all-purpose white and rye flours) are virtually absent from the mature leaven, and conversely, predominant genera in the mature leaven are undetected in flour.
The pH of a flour suspension is 6.2 (F in Figure 1A). The initial decrease in pH (dashed line in Figure 1A) from 6.2 to 5.7 at time = 0 is due to the mixture of the mature leaven at a pH of 3.85 (see below) in the mass proportion of 1:20 with flour. Subsequently, the pH decreases more slowly during the culture’s first 2 h and more rapidly over the next 10 h due to the growth of lactic acid bacteria (Figure 1B and below). The pH plateaus at 3.85 after 12 h, close to the pkA of lactic acid that was also observed in the 12 h mature starter (M in Figure 1A). The pH of the 24 h sample drifts up, perhaps due to the LAB earlier reaching a stationary phase (Figure 1B).
The cell densities of the bacteria and fungi in the mature leaven used in the feeding are 2.3 × 109/mL and 2.5 × 107/mL, respectively (Figure 1B,C), more than four orders of magnitude higher than the respective endogenous bacterial and fungal cell densities of 8 × 104/mL and 2 × 103/mL in a 50% w/v flour suspension (values for flour were omitted from the plots to avoid compressing the ordinate). The mixture at time = 0 is thus dominated by the bacteria and fungi from the mature leaven, with predicted cell densities of approximately 1.1 × 108/mL and 1.25 × 106/mL, respectively, which is consistent with observed values (Figure 1B,C).
The bacteria display their maximal growth rate for the first four hours (Figure 1B) with a generation time of 85 min; the growth rate declines between four and eight hours and plateaus between eight and twenty-four hours. The fungi also display maximal exponential growth for the first four hours (Figure 1C) with a generation time of 106 min. The six-hour time point reproducibly falls below the smooth curve generated in R, suggesting a pause in fungal growth between four and eight hours.
The acidity of the sourdough starter limits the potential growth of foodborne pathogens (e.g., Listeria monocytogenes or Bacillus cereus). In a TSB medium that was adjusted to different pHs using HCl, the germination of B. cereus spores was largely inhibited at a pH below 4.2, while it was nearly unaffected at a pH above this value [19]. The culture pH crosses this pH boundary between six and eight hours (Figure 1A). The presence of lactate, a by-product of the lactic acid bacteria, limits spore germination by 50%. Listeria Monocytogenes is more resistant to acid shock in the conditions created by the lactic acid bacteria [20]. However, its lower limit is still a pH of 4.4–4.7. A decrease in the virulence gene transcription after 5 h at pH 4.0 was achieved with acetic acid, another fermentative byproduct [21]. These acids make the starter naturally resistant to strains of foodborne pathogens, including those with acid tolerance responses [22]. The same is broadly described for the bread spoilage fungi, including Aspergillus and Penicillium [23].
The colonial and microscopic appearance of abundant microorganisms. Distinctive colony morphologies were observed on the plates obtained from twelve-hour cultures (Figure 2). The wet mount micrographs display the distinguishing characteristics of the cells, which were also reliably identified to a genus level by Sanger sequencing (Figure 3).
The large fungal colonies (Figure 2A) are Pichia, a yeast that grows in short chains and small clusters (see micrograph, below left). Small fungal colonies are Saccharomyces, a budding yeast that grows individually. The proportion of plate colonies favors Pichia over Saccharomyces by about 4:1, which is comparable to the proportions observed by metabarcoding (see below and Section 4).
Two bacterial genera were observed in roughly equal proportions (Figure 2B), in approximate agreement with the metabarcoding results (see below); Lactiplantibacillus plantarum forms larger white colonies of longer rod-shaped cells than those observed in the smaller translucent colonies of Lactobacillus brevis (see also Figure 3).
The LAB assigned to distinct genera by PCR and Sanger sequencing. The DNAs extracted from the plate colonies (Figure 2B) were amplified using the 16S primer pair (Table 1) and sequenced using M13 forward and reverse primers (see Section 2), enabling an assembly of full-length ~1450 bp amplicons. The first two lines of NCBI-BLASTn query/subject alignments are shown (Figure 3A,B).
Lactiplantibacillus plantarum has a 3,194,447 bp genome (indicated by the dashed red ellipses) while Lactobacillus brevis has a 2,351,988 bp genome (red ellipses in Figure 3A,B). Both of the genera have five dispersed rRNA gene copies (ranges).
Lactobacillus taxonomy has been controversial for over a century [25]. Most recently, Lactiplantibacillus plantarum, often still referred to as Lactobacillus plantarum, was authoritatively classified as a separate genus [24]. The first two lines of a multiple sequence alignment (Figure 3C), within which seven SNPs and four dinucleotide indels occur within the first 120 bp of the alignment, accompanying the observed differences in colony morphology, cell size (Figure 2B) and genome size (Figure 3A,B), amply support the assignment of Lactiplantibacillus plantarum and Lactobacillus brevis as separate genera.
Metagenomic proportions observed in the 16S microbiome. The genera in the 16S microbiome of the starting material (Flour columns in Figure 4) are completely different from those toward the end of the time course (12 h, 24 h and Mature). Strikingly, the plant plastid DNAs (chloroplast and mitochondria; gray and violet, respectively) swamp out the bacteria in the flour sample. Fortunately, enough bacterial reads were obtained (approximately 3000 from the representative ASV Metadata.qzv file) to determine the proportions of the bacterial genera endogenous to the flour mixture.
As the culture was started at t = 0, the abundant plant plastid material was first diluted with the LAB (Figure 4), which was added with the mature leaven to the flour in a mass proportion of 1:20, but with approximately four orders of magnitude higher cell density. The proportion of the plant plastid material further decreases with additional culture time, as illustrated graphically in the superimposed plastid proportion plots (Figure 5), approaching zero after 6 h of culture, presumably due to consumption during fermentation.
The rapid supplanting of superabundant plant plastid material (Figure 5) by the LAB is further illustrated by the superimposed LAB proportion plots (Figure 6). The proportions of Lactiplantibacillus plantarum and Lactobacillus brevis remain roughly equal throughout the growth of the culture, virtually replacing the plant plastids and bacteria endogenous to flour. A third LAB genus, Lacticasseibacillus, persists at a proportion of 1% throughout the culture period (Figure 6). The decline in the proportion of plant plastids superimposed on the increasing combined proportion of the LAB suggests consumption and growth, respectively (Figure 7).
Bacterial genera endogenous to flour, first swamped out by the LAB that is added with the starter, are soon overwhelmed by the rapid growth of the LAB. The plant plastids were deleted from the microbial composition data to visualize the bacteria endogenous to flour (Figure 8), comparable in design to the 16S proportion plots, which included the plant plastids (Figure 4, Figure 5, Figure 6 and Figure 7). A log scale was used on the ordinate of the corresponding proportion plots (Figure 9) to visualize the rapid replacement of bacteria endogenous to flour by the LAB. Fifteen bacterial genera were identified in the flour with the exclusion of plant plastids (legend at the bottom of Figure 8). The top four genera (Pantoea, Curtobacterium, Sphingomonas and Pseudomonas) together constitute 72% of the total bacterial genera; 11 less prominent genera constitute the remaining 28% (Figure 9, cf Figure 8). The LAB are practically undetectable in flour, and the bacteria endogenous to flour are negligible after outgrowth of the LAB during the time course.
The fungal microbiome changes more gradually with the growth of the leaven. Alternaria and Blumeria are the most abundant fungal genera endogenous to flour, while Pichia and Saccharomyces, the yeast genera that dominate mature leaven, were not detected in the flour (Figure 10, Figure 11 and Figure 12). The Alternaria proportionate decay curve (Figure 11) is less steep than that of the plant plastids and the bacteria endogenous to flour; Alternaria (light green in the stacked bar plots; Figure 10) still represents a significant proportion of the fungal microbiome at 8 h, becoming negligible only in the 12 h culture (Figure 10 and Figure 11). Alternaria may better tolerate the acidification that arises from the outgrowth of the LAB compared to the other fungal and bacterial genera endogenous to flour. Pichia predominates over Saccharomyces, with a proportion that fluctuates between 0.8/0.2 and 0.9/0.1 from 6 to 24 h and in the mature culture (Figure 12).
Pichia primarily uses aerobic metabolism, while S. cerevisiae often switches to fermentation even when oxygen is available. The predominance of Pichia over the long term correlates with its more oxidative metabolism [26]. As a result of its respiratory metabolism, Pichia can be cultured to exceptionally high cell densities [19]. The proportional advantage of Pichia over Saccharomyces is only observed after 24 h when the fungal cell density in the culture starts to fall (Figure 12, cf Figure 1C), suggesting that the proportionate increase in Pichia over Saccharomyces arose over a longer term, including a series of 24 h feedings combined with weeks of the culture dormant under refrigeration.
A twelve-hour culture shown above, when prepared at home, was used to leaven bread following the Tartine recipe. The foregoing illustrates the importing of home-baking materials to the research lab to investigate the changes in the fermentative microbiome during the time course of leaven maturation. While this was underway, a 12 h expanded leaven incubated at home was used in the Tartine recipe [8], as illustrated in Figure 13. Walnuts and black raisins were added to the dough in the last fold before proofing.
The least variable microbial component of a leaven is the LAB, although some starters also prominently display AAB [1]. Variability is also observed in the fungal genera, and the predominance of Pichia in our culture, although unusual, is evidently tolerable.

4. Discussion

The microbiome studies could hardly be reported without metagenomics, but classical methods, including pH measurement, plating, micrography, Sanger sequencing and multiple sequence alignment, fill the gaps left by metagenomic analysis. Metagenomics covers the breadth of genera and their relative proportions. However, bacterial and fungal amplicon libraries must be analyzed separately, without a basis for inter-domain comparisons. Although imperfect, plating thus provides needed information on the relative densities of the bacteria and fungi and their growth rates.
The plate counts may not accurately reflect the microbial proportions, and the metagenomics can reveal previously unknown microbial taxa. However, in the present study, the predominant fungi (Pichia and Saccharomyces) and bacteria (Lactiplantibacillus plantarum and Lactobacillus brevis) are readily observed on plates in proportions broadly consistent with those obtained from metagenomic analyses (Figure 2, cf Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10, Figure 11 and Figure 12). The variation in colony growth rate is a minor challenge; Pichia colonies appear with overnight incubation at 27 °C, while Saccharomyces colonies only reach their maximal number after incubating for three days. The LAB colonies grew slowly, but comparable colony counts were observed after incubation at 37 °C for three and seven days.
The rRNA gene sequence and copy number variation complicate the estimation of microbial proportions. Since the invention of rapid DNA sequencing and PCR five four decades ago, respectively, these methods have been used for the identification of microbial taxa using the ribosomal RNA gene repeat for both fungi [11] and bacteria [10]. Sanger sequencing of PCR amplicons presents technical problems due to sequence variation between the repeats of a single colony clone, especially close to the 5′ end of the amplicon, as observed with Pantoea and various fungi. This challenge was circumvented with the design of a Pantoea-specific internal forward sequencing primer (Table 1, note 5) and more generally by metabarcoding, since single DNA molecules amplified in the separate clusters of an Illumina flow cell produce paired-end reads that are analogous to results that would be obtained by amplicon plasmid cloning.
The subject NCBI reports of bacterial whole chromosome sequences with a 16S amplicon sequence as queries include the number of matches (rRNA gene copy number), presented individually as ranges (Figure 3), interspersed over the circular bacterial chromosome. When comparing bacterial taxa with different numbers of matches, a correction factor would have to be introduced to convert from a metabarcoding report of microbial composition to a true proportion, which is not, however, required for Lactiplantibaccilus plantarum and Lactobacillus brevis, with five matches each. Furthermore, relative plate counts based on differences in appearance of the colonies (Figure 2B) broadly agree with the approximately 50/50 proportions obtained by metagenomic analysis (Figure 4 and Figure 8).
The fungal rRNA gene copy numbers vary greatly both within and between taxa. Publicly available whole genome sequences have been analyzed using a method that does not require an annotated reference genome [27] and, more recently, using a modal frequency method specifically for S. cerevisiae strains [28], for which a fully annotated reference genome is available. The variation in rRNA gene copy number in S. cerevisiae initially reported as 140 [29], cf [30], covers approximately an order of magnitude when numerous recorded S. cerevisiae whole genome sequences were compared (see the violin plots in [22]), while the rRNA gene copy number between reported fungal taxa varies over more than two orders of magnitude, from ~10 to over 1000 [27]. A scaling factor would therefore need to be introduced to convert the amplicon observational taxonomic units (OTUs) so as to correct for variation in the number of rRNA gene repeats in fungal genera. This value would have to be experimentally determined for a statistically valid number of independent isolates of each genus present in a natural mixture, which is beyond the scope of this project. The rRNA gene copy number in Pichia is reportedly ~35 [31], suggesting that a conversion factor of four would have to be introduced to roughly normalize Pichia to Saccharomyces OTUs. The colony proportions (based on the colony count and size, microscopy and Sanger sequencing; Figure 2A) do not, however, strikingly differ from the raw fungal OTUs shown in the stacked bar plots (Figure 10) and proportion plots.
Microbial composition of leaven microbiomes. The LAB generally prevail in leaven bacterial microbiomes, although acetic acid bacteria (AAB) such as Acetobacter have also been observed [1,2]. We found an equal proportion of the LAB Lactiplantibacillus plantarum and Lactobacillus brevis in the mature and 24 h cultures. Our fungal microbiome is presently dominated by Pichia, accompanied by Saccharomyces.
The rapid decrease in proportion of plant biological material during leaven maturation (Figure 4, Figure 5 and Figure 7) and domination of the mature leaven by two LAB and two fungal taxa (Figure 4, Figure 6, Figure 7, Figure 10 and Figure 11) illustrate both the consumption and growth characteristics of fermentation. Starting and end points have been widely reported [1,2,7]. The changes in microbial composition over 24 h documented here were enabled by the twenty-fold feeding of the culture (following [8]), which allowed cell density to increase by more than an order of magnitude, including several hours of exponential growth. This is not a feature of the two-fold feeding (back-slopping) used by others.
The microbiomes typical of flour and mature leaven are mutually exclusive. The endogenous flour microbiome, replenished upon feeding, is at first overwhelmed by the microbiome introduced with the mature starter and then is extinguished over the ensuing 12 h by rapid outgrowth of the starter microbiome.

5. Conclusions

The interest in natural fermentations is now widely accompanied by laboratory microbiome analysis. Following the growth of a leaven over twenty-four hours, we observed the rapid disappearance of the bacteria and fungi endogenous to flour and of plant plastid DNAs, and the exponential propagation of bacteria and fungi introduced by feeding with mature leaven. No overlap was observed between microbiomes endogenous to flour and mature leaven, which is consistent with the introduction of fermentative taxa from the exogenous environment. The mature leaven microbiome replaces the microbiome endogenous to flour in every growth cycle.

Author Contributions

Conceptualization, L.L. and J.G.; methodology, L.L. and M.D.; software, M.S. and R.M.; validation, M.S. and R.M.; formal analysis, L.L., M.S. and R.M.; investigation, L.L., M.S., J.G. and M.D.; resources, L.L.; data curation, M.S. and R.M.; writing—original draft preparation, L.L.; writing—review and editing, M.S., J.G., M.D., R.M. and L.L.; visualization, L.L. and M.S.; supervision, L.L. and R.M.; project administration, L.L.; funding acquisition, L.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by USDE grant numbers P120A220050 and P120A210053 and by The Pinkerton Foundation grant number G-2509-23017. The APC was funded by USDE grant number P120A210053.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data supporting reported results can be found at https://github.com/MonishaSherpa/Sourdough_Leaven_Metagenomics (accessed on 17 March 2026). The SeqCenter Fastq paired-end read files can be accessed at the NCBI SRA website PRJNA1426134, accessed on 21 March 2026.

Acknowledgments

We acknowledge the helpful discussions with colleagues at York College of The City University of New York, including Louis Bradbury, Monica Mowery, Melvin Silberklang and Ivica Arsov (Department of Biology), Emmanuel Chang (Department of Chemistry) and Derek Brazill. STEM-Care and STEM Academy Workshops sponsored by US Department of Education Grant #s P120A220050 and P120A210053 to The City University of New York, respectively, provided stipends and support for materials and services. O. Feliciano provided technical support. Karen Alban was a Stem-Care participant in the early stages of the project. High school students participated during the academic year in the Urban Barcode Research Program and in summer STEM Research Mentoring Consortium Workshops sponsored by The Pinkerton Foundation, which also provided material support. Jeremy Seto (Department of Biology, New York City Technical College of CUNY) and personnel at DNALC-Brooklyn DNA Learning Center NYC at City Tech provided instructional workshops, material support and helpful discussions.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Landis, E.A.; Oliverio, A.M.; McKenney, E.A.; Nichols, L.M.; Kfoury, N.; Biango-Daniels, M.; Shell, L.K.; Madden, A.A.; Shapiro, L.; Wolfe, B.E. The diversity and function of sourdough starter microbiomes. eLife 2021, 10, e61644. [Google Scholar] [CrossRef]
  2. McKenney, E.A.; Nichols, L.M.; Alvarado, S.; Hardy, S.; Kemp, K.; Polmanteer, R.; Shoemaker, A.; Dunn, R.R. Sourdough starters exhibit similar succession patterns but develop flour-specific climax communities. PeerJ 2023, 11, e16163. [Google Scholar] [CrossRef]
  3. Alfonzo, A.; Miceli, C.; Nasca, A.; Franciosi, E.; Ventimiglia, G.; Di Gerlando, R.; Tuohy, K.; Francesca, N.; Moschetti, G.; Settanni, L. Monitoring of wheat lactic acid bacteria from the field until the first step of dough fermentation. Food Microbiol. 2017, 62, 256–269. [Google Scholar] [CrossRef]
  4. Clark, C.S.; Ohstrom, A.; Rolon, M.L.; Smith, M.; Wolfe, B.E.; Wee, J.; Van Buiten, C.B. Sourdough starter culture microbiomes influence physical and chemical properties of wheat bread. J. Food Sci. 2024, 89, 1414–1427. [Google Scholar] [CrossRef]
  5. Ponomarova, O.; Gabrielli, N.; Sévin, D.C.; Mülleder, M.; Zirngibl, K.; Bulyha, K.; Andrejev, S.; Kafkia, E.; Typas, A.; Patil, K.R.; et al. Yeast creates a niche for symbiotic lactic acid bacteria through nitrogen overflow. Cell Syst. 2017, 5, 345–357. [Google Scholar] [CrossRef]
  6. Margulis, L.; Bermudes, D. Symbiosis as a mechanism of evolution: Status of cell symbiosis theory. Symbiosis 1984, 1, 101–124. [Google Scholar]
  7. von Gastrow, L.; Michel, E.; Legrand, J.; Amelot, R.; Segond, D.; Guezenec, S.; Rué, O.; Chable, V.; Goldringer, I.; Sicard, D.; et al. Microbial community dispersal from wheat grains to sourdoughs: A contribution of participatory research. Mol. Ecol. 2023, 32, 2413–2427. [Google Scholar] [CrossRef] [PubMed]
  8. Robertson, C. Tartine Bread. Chronicle Books. 2013. Available online: https://cooking.nytimes.com/recipes/1016277-tartines-country-bread (accessed on 18 June 2013).
  9. Gilbert, G.S.; Diaz, A.; Bregoff, H.A. Seed disinfestation practices to control seed-borne fungi and bacteria in home production of sprouts. Foods 2023, 12, 747. [Google Scholar] [CrossRef]
  10. Weisburg, W.G.; Barns, S.M.; Pelletier, D.A.; Lane, D.J. 16S ribosomal DNA amplification for phylogenetic study. J. Bacteriol. 1991, 173, 697–703. [Google Scholar] [CrossRef] [PubMed]
  11. White, T.J.; Bruns, T.; Lee, S.J.W.T.; Taylor, J. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In PCR Protocols: A Guide to Methods and Applications; Academic Press, Inc.: New York, NY, USA, 1990; Volume 18, pp. 315–322. [Google Scholar]
  12. Altschul, S.F.; Gish, W.; Miller, W.; Myers, E.W.; Lipman, D.J. Basic local alignment search tool. J. Mol. Biol. 1990, 215, 403–410. [Google Scholar] [CrossRef] [PubMed]
  13. Edgar, R.C. MUSCLE: Multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004, 32, 1792–1797. [Google Scholar] [CrossRef]
  14. Bolyen, E.; Rideout, J.R.; Dillon, M.R.; Bokulich, N.A.; Abnet, C.C.; Al-Ghalith, G.A.; Alexander, H.; Alm, E.J.; Arumugam, M.; Asnicar, F.; et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 2019, 37, 852–857. [Google Scholar] [CrossRef]
  15. Marcel Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 2011, 17, 10–12. [Google Scholar] [CrossRef]
  16. Estaki, M.; Jiang, L.; Bokulich, N.A.; McDonald, D.; González, A.; Kosciolek, T.; Martino, C.; Zhu, Q.; Birmingham, A.; Knight, R.; et al. QIIME 2 enables comprehensive end-to-end analysis of diverse microbiome data and comparative studies with publicly available data. Curr. Protoc. Bioinform. 2020, 70, e100. [Google Scholar] [CrossRef]
  17. Hu, X.; Liu, S.; Li, E. Microbial community succession and its correlation with the dynamics of flavor compound profiles in naturally fermented stinky tofu. Food Chem. 2023, 427, 136742. [Google Scholar] [CrossRef]
  18. Ahmad, M.; Hirz, M.; Pichler, H.; Schwab, H. Protein expression in Pichia pastoris: Recent achievements and perspectives for heterologous protein production. Appl. Microbiol. Biotechnol. 2014, 98, 5301–5317. [Google Scholar] [CrossRef] [PubMed]
  19. Darby, R.A.; Cartwright, S.P.; Dilworth, M.V.; Bill, R.M. Which yeast species shall I choose? Saccharomyces cerevisiae versus Pichia pastoris (Review). In Recombinant Protein Production in Yeast: Methods and Protocols; Humana Press: Totowa, NJ, USA, 2012; pp. 11–23. [Google Scholar]
  20. Neuhaus, K.; Satorhelyi, P.; Schauer, K.; Scherer, S.; Fuchs, T.M. Acid shock of Listeria monocytogenes at low environmental temperatures induces prfA, epithelial cell invasion, and lethality towards Caenorhabditis elegans. BMC Genom. 2013, 14, 285. [Google Scholar] [CrossRef] [PubMed]
  21. Rieu, A.; Guzzo, J.; Piveteau, P. Sensitivity to acetic acid, ability to colonize abiotic surfaces and virulence potential of Listeria monocytogenes EGD-e after incubation on parsley leaves. J. Appl. Microbiol. 2010, 108, 560–570. [Google Scholar] [CrossRef] [PubMed]
  22. Shabala, L.; Lee, S.H.; Cannesson, P.; Ross, T. Acid and NaCl limits to growth of Listeria monocytogenes and influence of sequence of inimical acid and NaCl levels on inactivation kinetics. J. Food Prot. 2008, 71, 1169–1177. [Google Scholar] [CrossRef]
  23. Liu, A.; Xu, R.; Zhang, S.; Wang, Y.; Hu, B.; Ao, X.; Li, Q.; Li, J.; Hu, K.; Liu, S.; et al. Antifungal mechanisms and application of lactic acid bacteria in bakery products: A review. Front. Microbiol. 2022, 13, 924398. [Google Scholar] [CrossRef]
  24. Zheng, J.; Wittouck, S.; Salvetti, E.; Franz, C.M.; Harris, H.M.; Mattarelli, P.; O’Toole, P.W.; Pot, B.; Vandamme, P.; Lebeer, S.; et al. A taxonomic note on the genus Lactobacillus: Description of 23 novel genera, emended description of the genus Lactobacillus Beijerinck 1901, and union of Lactobacillaceae and Leuconostocaceae. Int. J. Syst. Evol. Microbiol. 2020, 70, 2782–2858. [Google Scholar] [CrossRef]
  25. White, B.; Avery, O.T. Observations on Certain Lactic Acid Bacteria of the So-Called Bulgaricus Type. Zbl. Bakt. Abt. 1910, 25, 161–184. [Google Scholar]
  26. Fiaux, J.; Çakar, Z.P.; Sonderegger, M.; Wüthrich, K.; Szyperski, T.; Sauer, U. Metabolic-flux profiling of the yeasts Saccharomyces cerevisiae and Pichia stipitis. Eukaryot. Cell 2003, 2, 170–180. [Google Scholar] [CrossRef] [PubMed]
  27. Lofgren, L.A.; Uehling, J.K.; Branco, S.; Bruns, T.D.; Martin, F.; Kennedy, P.G. Genome-based estimates of fungal rDNA copy number variation across phylogenetic scales and ecological lifestyles. Mol. Ecol. 2019, 28, 721–730. [Google Scholar] [CrossRef]
  28. Sharma, D.; Hermann-Le Denmat, S.; Matzke, N.J.; Hannan, K.; Hannan, R.D.; O’Sullivan, J.M.; Ganley, A.R. A new method for determining ribosomal DNA copy number shows differences between Saccharomyces cerevisiae populations. Genomics 2022, 114, 110430. [Google Scholar] [CrossRef]
  29. Goffeau, A.; Barrell, B.G.; Bussey, H.; Davis, R.W.; Dujon, B.; Feldmann, H.; Galibert, F.; Hoheisel, J.D.; Jaco, C.; Oliver, S.G.; et al. Life with 6000 genes. Science 1996, 274, 546–567. [Google Scholar] [CrossRef] [PubMed]
  30. Johnston, M.; Hillier, L.; Riles, L.; Genome Sequencing Center; Albermann, K.; André, B.; Ansorge, W.; Benes, V.; Brückner, M.; Hoheisel, J.D.; et al. The nucleotide sequence of Saccharomyces cerevisiae chromosome XII. Nature 1997, 387, 87–90. [Google Scholar] [CrossRef] [PubMed]
  31. Love, K.R.; Shah, K.A.; Whittaker, C.A.; Wu, J.; Bartlett, M.C.; Ma, D.; Leeson, R.L.; Priest, M.; Borowsky, J.; Love, J.C.; et al. Comparative genomics and transcriptomics of Pichia pastoris. BMC Genom. 2016, 17, 550. [Google Scholar] [CrossRef]
Fermentation 12 00171 g001
Figure 1. pH and cell density as a function of culture time. (A) Suspensions were diluted with water for pH readings (see Section 2). F and M, connected with dashed lines, are the flour mixture and the mature starter used to start the culture, respectively. (B,C) A series of ten-fold dilutions was made with sterile saline, and 100 μL of each dilution was plated. See Section 2 for details. Generation time is 85 min for the bacteria, based on the densities from 0 to 4 h, and 106 min for the fungi, based on 0–2 h.
Figure 1. pH and cell density as a function of culture time. (A) Suspensions were diluted with water for pH readings (see Section 2). F and M, connected with dashed lines, are the flour mixture and the mature starter used to start the culture, respectively. (B,C) A series of ten-fold dilutions was made with sterile saline, and 100 μL of each dilution was plated. See Section 2 for details. Generation time is 85 min for the bacteria, based on the densities from 0 to 4 h, and 106 min for the fungi, based on 0–2 h.
Fermentation 12 00171 g001
Fermentation 12 00171 g002
Figure 2. The plate images and wet mount micrographs of (A) fungi and (B) bacteria obtained from the 12 h starter culture. The large fungal colonies are Pichia; the small fungal colonies are Saccharomyces. The large whiter bacterial colonies are Lactiplantibaccilus plantarum; the small greyer bacterial colonies are Lactobaccilus brevis. The colonies (red circles) were picked with sterile tips and suspended in sterile saline for micrography. The wet mount micrographs of fungi at 400×; the bacteria at 1000× (oil immersion). The hemocytometer slide was used to provide a dimensional marker (50 μm) and to ease finding the focal plane.
Figure 2. The plate images and wet mount micrographs of (A) fungi and (B) bacteria obtained from the 12 h starter culture. The large fungal colonies are Pichia; the small fungal colonies are Saccharomyces. The large whiter bacterial colonies are Lactiplantibaccilus plantarum; the small greyer bacterial colonies are Lactobaccilus brevis. The colonies (red circles) were picked with sterile tips and suspended in sterile saline for micrography. The wet mount micrographs of fungi at 400×; the bacteria at 1000× (oil immersion). The hemocytometer slide was used to provide a dimensional marker (50 μm) and to ease finding the focal plane.
Fermentation 12 00171 g002
Fermentation 12 00171 g003
Figure 3. Assignment of lactic acid bacteria to distinct genera based on Sanger sequencing. Here, 16S amplicons were amplified from DNA extracted from colonies (see Section 2) and sequenced by Genewiz with M13 F and R primers (see Section 2). Full-length contigs were assembled, and BLASTn analysis identified (A) the genus Lactiplantibacillus plantarum from the larger, whiter colonies (see Figure 2) and (B) Lactobacillus brevis from the smaller, greyer colonies. The dashed red ellipses highlight contig length and sequence quality, genera, very different genome sizes, and ranges (rRNA gene copy number) (C) The first two lines of a multiple sequence (EMBL MUSCLE) show numerous single and double nucleotide substitutions and deletions, consistent with assignment of different genera (see [24] for a detailed and comprehensive treatment of the topic). The top two sequences in the alignment, B9 and B10, are Lactobacillus brevis; the bottom two sequences, B7 and B8, are Lactiplantibacillus plantarum. *s below the alignments indicate identity of all four isolates.
Figure 3. Assignment of lactic acid bacteria to distinct genera based on Sanger sequencing. Here, 16S amplicons were amplified from DNA extracted from colonies (see Section 2) and sequenced by Genewiz with M13 F and R primers (see Section 2). Full-length contigs were assembled, and BLASTn analysis identified (A) the genus Lactiplantibacillus plantarum from the larger, whiter colonies (see Figure 2) and (B) Lactobacillus brevis from the smaller, greyer colonies. The dashed red ellipses highlight contig length and sequence quality, genera, very different genome sizes, and ranges (rRNA gene copy number) (C) The first two lines of a multiple sequence (EMBL MUSCLE) show numerous single and double nucleotide substitutions and deletions, consistent with assignment of different genera (see [24] for a detailed and comprehensive treatment of the topic). The top two sequences in the alignment, B9 and B10, are Lactobacillus brevis; the bottom two sequences, B7 and B8, are Lactiplantibacillus plantarum. *s below the alignments indicate identity of all four isolates.
Fermentation 12 00171 g003
Fermentation 12 00171 g004
Figure 4. Stacked bar plots for 16S amplicon proportions during the time course. Triplicate samples of flour are dominated by plant plastids (chloroplast (gray) and mitochondria (violet)). After feeding with starter at time = 0_h, the plant plastids are first diluted with lactic acid bacteria from the starter, and a subsequent decrease in the proportion of plant plastids (2_h to 24_h) indicates the rapid consumption of the plant material as food for the microorganisms, approaching a steady state within 8 h of culture. M, on the right, is the mature sample used to start the time course by mixing with water and flour (see Section 2).
Figure 4. Stacked bar plots for 16S amplicon proportions during the time course. Triplicate samples of flour are dominated by plant plastids (chloroplast (gray) and mitochondria (violet)). After feeding with starter at time = 0_h, the plant plastids are first diluted with lactic acid bacteria from the starter, and a subsequent decrease in the proportion of plant plastids (2_h to 24_h) indicates the rapid consumption of the plant material as food for the microorganisms, approaching a steady state within 8 h of culture. M, on the right, is the mature sample used to start the time course by mixing with water and flour (see Section 2).
Fermentation 12 00171 g004
Fermentation 12 00171 g005
Figure 5. Decay in the proportion of plant plastids during the time course. The proportion of plastids approaches zero after 8 h of culture. Designations F and M flanking the culture time course are as in Figure 4.
Figure 5. Decay in the proportion of plant plastids during the time course. The proportion of plastids approaches zero after 8 h of culture. Designations F and M flanking the culture time course are as in Figure 4.
Fermentation 12 00171 g005
Fermentation 12 00171 g006
Figure 6. Growth in the proportion of lactic acid bacteria (LAB) from the addition of mature starter at time = 0. Three genera and the sum of the three LAB shown were plotted as labeled. Initially, Lactiplantibacillus grows faster than Lactobacillus and its proportion plateaus at 4 h culture time; the proportion of Lactobacillus rises more slowly and plateaus at 8 h. Thereafter, these two genera hold roughly steady proportions. Lacticasseibacillus remains at a low proportion throughout the culture time.
Figure 6. Growth in the proportion of lactic acid bacteria (LAB) from the addition of mature starter at time = 0. Three genera and the sum of the three LAB shown were plotted as labeled. Initially, Lactiplantibacillus grows faster than Lactobacillus and its proportion plateaus at 4 h culture time; the proportion of Lactobacillus rises more slowly and plateaus at 8 h. Thereafter, these two genera hold roughly steady proportions. Lacticasseibacillus remains at a low proportion throughout the culture time.
Fermentation 12 00171 g006
Fermentation 12 00171 g007
Figure 7. Superimposed proportion plots for the LAB and plant plastids. The proportion of the total LAB increases, evidently at the expense of decreasing plant plastids.
Figure 7. Superimposed proportion plots for the LAB and plant plastids. The proportion of the total LAB increases, evidently at the expense of decreasing plant plastids.
Fermentation 12 00171 g007
Fermentation 12 00171 g008
Figure 8. Bacterial stacked bar plots after the removal of plant plastids (from the data used to generate Figure 4). The highest proportion of the bacteria present in the flour mixture, Pantoea, Curtobacterium, Sphingomonas and Pseudomonas, are overshadowed by the lactic acid bacteria by mixing with the starter in the time = 0 sample, followed by further growth of the LAB.
Figure 8. Bacterial stacked bar plots after the removal of plant plastids (from the data used to generate Figure 4). The highest proportion of the bacteria present in the flour mixture, Pantoea, Curtobacterium, Sphingomonas and Pseudomonas, are overshadowed by the lactic acid bacteria by mixing with the starter in the time = 0 sample, followed by further growth of the LAB.
Fermentation 12 00171 g008
Fermentation 12 00171 g009
Figure 9. Proportion decay curves for Pantoea, Curtobacterium, Sphingomonas and Pseudomonas, the four most abundant bacterial taxa and the sum of these top four with plant plastids removed from the table. The ordinate is on a log scale. The abundant bacterial taxa in the flour rapidly decrease in proportion in the culture, first by mixing with the starter at time = 0, and subsequently due to the growth of LAB (as seen in Figure 7 and Figure 8). Dashed lines from F to 0 indicate the proportional decrease due to mixing.
Figure 9. Proportion decay curves for Pantoea, Curtobacterium, Sphingomonas and Pseudomonas, the four most abundant bacterial taxa and the sum of these top four with plant plastids removed from the table. The ordinate is on a log scale. The abundant bacterial taxa in the flour rapidly decrease in proportion in the culture, first by mixing with the starter at time = 0, and subsequently due to the growth of LAB (as seen in Figure 7 and Figure 8). Dashed lines from F to 0 indicate the proportional decrease due to mixing.
Fermentation 12 00171 g009
Fermentation 12 00171 g010
Figure 10. Fungal stacked bar plots. Alternaria, abundant in flour, decreases in proportion while Pichia and Saccharomyces, introduced with the starter, increase in proportion to dominate the culture between 8 and 12 h.
Figure 10. Fungal stacked bar plots. Alternaria, abundant in flour, decreases in proportion while Pichia and Saccharomyces, introduced with the starter, increase in proportion to dominate the culture between 8 and 12 h.
Fermentation 12 00171 g010
Fermentation 12 00171 g011
Figure 11. Alternaria proportional decay curve. Alternaria, introduced with flour, decreases over culture time, but not as quickly as the endogenous bacteria (see Figure 9).
Figure 11. Alternaria proportional decay curve. Alternaria, introduced with flour, decreases over culture time, but not as quickly as the endogenous bacteria (see Figure 9).
Fermentation 12 00171 g011
Fermentation 12 00171 g012
Figure 12. Proportion plot for Pichia and Saccharomyces. Pichia is abundant throughout all culture times, with Saccharomyces maintaining a steady state between 15 and 20, and the combined proportion of Pichia and Saccharomyces never reaches 100% due to a low level of persistence of Alternaria (see Figure 11).
Figure 12. Proportion plot for Pichia and Saccharomyces. Pichia is abundant throughout all culture times, with Saccharomyces maintaining a steady state between 15 and 20, and the combined proportion of Pichia and Saccharomyces never reaches 100% due to a low level of persistence of Alternaria (see Figure 11).
Fermentation 12 00171 g012
Fermentation 12 00171 g013
Figure 13. Tartine bread with raisins and walnuts added during the last fold before proofing. The leaven was prepared from a refrigerator stock by feeding twice at 12 h intervals.
Figure 13. Tartine bread with raisins and walnuts added during the last fold before proofing. The leaven was prepared from a refrigerator stock by feeding twice at 12 h intervals.
Fermentation 12 00171 g013
Table 1. The oligonucleotides used for 16S and ITS amplification and sequencing.
Table 1. The oligonucleotides used for 16S and ITS amplification and sequencing.
NameSequence 5′ to 3′Notes
16S-FTGTAAAACGACGGCCAGTAGAGTTTGGATCMTGGCTCAG1
16S-RCAGGAAACAGCTATGACCGGTTACCTTGTTACGACTT2
ITS-FTGTAAAACGACGGCCAGTCCGTAGGTGAACCTGCGG3
ITS-RCAGGAAACAGCTATGACTCCTCCGCTTATTGATATGC4
PantoeaCTTGGGTGACGAGTGGC5
The amplicon sequence is in bold; the sequence on the 5′ side is the M13 sequencing tag. 1,2 fD1 and rP2, respectively [10]. 3,4 The ITS1F and ITS4R universal primers [11]. 5 This study; an internal 5′ end Pantoea-specific sequencing primer corresponding, for example, to nt 38–54 in the 1415 bp Pantoea agglomerans strain 29bL4A 16S ribosomal RNA gene, partial sequence, Sequence ID: MT878369.1.
Table 2. The SeqCenter oligonucleotides for the V3/V4 16S (bacterial) and ITS2 (fungal) amplicon library preparation.
Table 2. The SeqCenter oligonucleotides for the V3/V4 16S (bacterial) and ITS2 (fungal) amplicon library preparation.
NameSequence 5′ to 3′Notes
V3/V4 341fCCTACGGGDGGCWGCAG, CCTAYGGGGYGCWGCAG1
V3/V4 806rGACTACNVGGGTMTCTAATCC1
ITS3fGCATCGATGAAGAACGCAG2
ITS4rTCCTCCGCTTATTGATATGC2
1 From https://www.seqcenter.com/service/metagenome-sequencing/16s-its-sequencing/ (accessed on 14 December 2023) (plant plastid sequences included in library). 2 From https://www.seqcenter.com/service/metagenome-sequencing/16s-its-sequencing/ (accessed on 14 December 2023).
Table 3. The student participation.
Table 3. The student participation.
ProgramTypeContact HrsPeriod# Part.
STEM-CARE 1UG120/SemF’23–Sp’252
UBRP 2HS 555/2 SemF’24–Sp’262
WSAW 3UG36/3 WksJan’24, Jan’2515
SRMC 4HS 572/6 WksJuly–Aug, ’24, ’255
SSAW 3UG36/3 WksJuly–Aug, ’2515 *
1 the Science, Technology, Engineering, Mathematics, Career Advisement and Research Engagement programs were sponsored by the U.S. Department of Education Minority Science and Engineering Improvement Program grant to CUNY. 2 Urban Barcoding Research Program, an academic year internship program for NYC high school students funded by The Pinkerton Foundation. 3 winter and summer STEM Academy Workshops funded by the U.S. Department of Education. 4 STEM Research Mentorship Consortium, a summer internship program for NYC high school students funded by The Pinkerton Foundation. 5 the participating high schools included Queens High School for the Sciences at York College, York Early College Academy, Hillcrest High School, Manhattan Center for Science and Mathematics, and John Adams High School. * Undergraduates joined the high school students for the second half of the summer 2025 workshop.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Levinger, L.; Sherpa, M.; Gelman, J.; Dibonaventura, M.; Mandal, R. A Day in the Life of a Sourdough Leaven from Feeding to Maturity. Fermentation 2026, 12, 171. https://doi.org/10.3390/fermentation12040171

AMA Style

Levinger L, Sherpa M, Gelman J, Dibonaventura M, Mandal R. A Day in the Life of a Sourdough Leaven from Feeding to Maturity. Fermentation. 2026; 12(4):171. https://doi.org/10.3390/fermentation12040171

Chicago/Turabian Style

Levinger, Louis, Monisha Sherpa, Julia Gelman, Mariapia Dibonaventura, and Rabindra Mandal. 2026. "A Day in the Life of a Sourdough Leaven from Feeding to Maturity" Fermentation 12, no. 4: 171. https://doi.org/10.3390/fermentation12040171

APA Style

Levinger, L., Sherpa, M., Gelman, J., Dibonaventura, M., & Mandal, R. (2026). A Day in the Life of a Sourdough Leaven from Feeding to Maturity. Fermentation, 12(4), 171. https://doi.org/10.3390/fermentation12040171

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop