Development and implementation of a project-based laboratory module: heterologous expression and fermentation of ergothioneine in E. coli for undergraduate biochemistry and molecular biology education

ABSTRACT

There is a growing need in biochemistry education to move beyond traditional, recipe-based experiments and toward integrated, research-based experiences that mimic real-world scientific inquiry. We designed a multi-week, project-based learning (PBL) module where students undertake the complete process of producing ergothioneine (ERG), a valuable antioxidant, using recombinant DNA technology in Escherichia coli. The project encompasses molecular cloning (optional), protein expression analysis, fermentation in a bioreactor, and product quantification. The module was successfully implemented in an upper-level biochemistry laboratory course. Assessment via pre- and post-lab quizzes, lab reports, and a comprehensive student self-evaluation survey showed significant gains in technical skills, conceptual understanding, and research confidence. This laboratory module provides an effective framework for integrating modern biotechnological techniques into the undergraduate curriculum. It enhances student learning by contextualizing theoretical knowledge within a continuous, authentic research project, thereby fostering critical thinking, problem-solving skills, and enthusiasm for biochemical research.

INTRODUCTION

The limitations of traditional “cookbook” labs are well-documented (1–3). By having students follow step-by-step protocols to achieve predictable results, these exercises prioritize verification over genuine discovery. Students perform tasks mechanically, missing opportunities to develop critical skills in experimental design and problem-solving. This approach fosters a misconception of science as a linear pursuit of right answers, rather than an iterative process of inquiry (4, 5). Additionally, teaching techniques in isolation prevents students from understanding their application in a broader research context. The lack of authentic engagement, particularly with unexpected or negative data, diminishes student motivation and fails to prepare them for the complexities of real-world scientific investigation (6, 7).

Given these limitations, educators have sought alternative pedagogical approaches that better engage students in authentic scientific practices. Project-based learning (PBL) offers a powerful alternative by situating students within an extended, authentic research process (8–10). Unlike cookbook labs, PBL requires students to design experiments, respond to unexpected results, and troubleshoot technical challenges, thereby cultivating resilience, analytical skills, and a sense of ownership over their work. Collaboration within research teams mirrors real scientific environments, strengthening interpersonal and communication skills. Ultimately, PBL promotes deeper conceptual understanding and empowers students to “think like scientists,” better preparing them for careers in research, industry, and medicine (11–14).

However, the successful implementation of PBL depends critically on selecting an appropriate research theme, one that is both scientifically rich and pedagogically tractable. Ergothioneine (ERG) biosynthesis in Escherichia coli represents an ideal project theme for several reasons. First, ERG, a sulfur-containing antioxidant known as the “longevity vitamin,” holds significant commercial value for mitigating oxidative stress and age-related diseases (15), providing students with a real-world context. Second, the well-established ERG biosynthesis pathways, particularly the bacterial egtA/B/C/D/E cluster (Mycobacterium smegmatis) and fungal Egt1/Egt2 system (Neurospora crassa) (Fig. 1), offer a genuine engineering challenge (16–19). Students compare these pathways and then design a synthetic biology solution by selecting and justifying the expression of prokaryotic, eukaryotic, or hybrid gene clusters in E. coli. Third, the flexibility and amenability of these genes to expression in E. coli thus transform the microbe into a living laboratory, making ERG biosynthesis a powerful platform for immersive, project-based learning in synthetic biology and fermentation science. Through this process, students engage directly with core concepts of heterologous expression, pathway optimization, and metabolic burden.

Fig 1.

Pathway of ergothioneine biosynthesis. SAM: S-adenosylmethionine; γ-GC: γ-glutamyl-cysteine; HER: hercynine; γGC-HER: hercynyl-γ-glutamyl-cysteine sulfoxide; Cys-HER: hercynyl-cysteine sulfoxide; MetK: S-adenosylmethionine synthetase. Black represents two or more public biosynthetic pathways. Red represents biosynthetic pathways of ERG in M. smegmatis (EgtA: γ-glutamyl cysteine synthase; EgtB: mononuclear non-heme iron enzyme; EgtC: amidotransferase; EgtD: SAM-dependent histidine methyltransferase; EgtE: PLP-mediated C-S lyase). Purple represents biosynthetic pathways of ERG in N. crassa (Egt1: Bifunctional enzymes [SAM-dependent histidine methyltransferase and mononuclear nonheme iron enzyme]; Egt2: PLP-mediated C-S lyase). Orange represents biosynthetic pathways of ERG in Methylobacterium (EgtB: Similar to fungi Egt1). Pink represents biosynthetic pathways of ERG in anaerobic bacteria (EanA: Methyltransferase; EanB: rhodanese-like sulfur transferase). Green represents the participation of L-Glu in the M. smegmatis biosynthetic pathway.

Based on the pedagogical rationale and scientific suitability outlined above, we designed a teaching module that uses PBL to guide students through the full workflow of ERG production and quantification in E. coli (Fig. 2). The specific educational goals of this module are aligned with project activities and assessment methods, as summarized in Table S1. These goals include: (i) providing an authentic research experience covering the entire bioprocess from genetic design to product analysis; (ii) reinforcing core concepts in biochemistry, molecular biology, and microbial physiology through hands-on work; (iii) building essential research skills including experimental design, technique, data analysis, problem-solving, and scientific communication; and (iv) increasing student engagement and career readiness through a challenging, collaborative project with real-world relevance.

Fig 2.

Integrated framework of an undergraduate laboratory module: timeline, teaching methodology, and assessment methods

PROCEDURE

Learning context and course design

This project-based learning (PBL) module was designed for upper-level undergraduate students (5th–6th semester, Year 3) majoring in Bioengineering at Hebei University of Technology. The module served as a required component of an advanced laboratory course, with the central project goal of optimizing ergothioneine (ERG) production through heterologous gene expression.

The module aimed to develop three core competencies: (i) foundational knowledge of heterologous gene expression and metabolic engineering principles, (ii) proficiency in basic molecular biology techniques (bacterial transformation, PCR, gel electrophoresis, SDS-PAGE, and HPLC), and (iii) the ability to design experiments, analyze data, and communicate scientific findings in written and oral formats.

The 8-week module employed a hybrid online-offline model. Unlike traditional step-by-step laboratory instruction, students worked in teams of 3–4 to conduct literature reviews, generate hypotheses, select methods, and troubleshoot experiments, mirroring the authentic research process. Instructors served as facilitators, guiding inquiry through Socratic questioning rather than dictating procedures.

To successfully engage with this research-oriented project, students were required to have completed foundational coursework in Biochemistry, Molecular Biology, and Microbiology and demonstrated practical experience with bacterial transformation, PCR, and gel electrophoresis.

Research design and participant selection

Study design

This study employed a quasi-experimental, two-cohort design to evaluate the effectiveness of the PBL approach against a traditional laboratory instruction model. The control cohort (Class of 2022, n = 30) received traditional step-by-step laboratory instruction with prescriptive protocols. The experimental cohort (Class of 2023, n = 32) completed the PBL module described above, working on the same ERG production project but without detailed protocols.

To ensure comparability, strict standardization was maintained across both cohorts. The core learning goals and final assessment criteria (technical skills rubric and knowledge-based examination) were identical for both groups. All participants provided informed consent, and the study received ethical approval from the university’s institutional review board (Approval No: HEBUT-2023-BCHE-028).

Participant characteristics

A total of 62 junior bioengineering students participated across 2 academic years (2022 and 2023). Participant characteristics are summarized in Table 1.

TABLE 1.

Demographic characteristics of participants^a

Characteristic	Control group (2022, n = 30)	Experimental group (2023, n = 32)
Gender (male/female)	16/14	17/15
Age (mean ± SD, years)	21.2 ± 0.8	21.3 ± 0.7
Age range (years)	20–23	20–23
Academic level	5th–6th semester (Year 3)	5th–6th semester (Year 3)
Major: Bioengineering	30	32

^a The control cohort initially consisted of 32 students, but two students withdrew prior to final assessment due to personal reasons, leaving a final n of 30. The experimental cohort consisted of 32 students, all of whom completed the program (n = 32).

Assessment of prior knowledge

To establish baseline equivalence, prior knowledge was assessed using the following three complementary measures: (i) overall cumulative GPA: A pre-class background assessment administered during the first week of the semester measured students’ cumulative Grade Point Averages. An independent samples t-test found no statistically significant difference between the control group (M = 3.18, SD = 0.42) and the PBL group (M = 3.19, SD = 0.41); t (60) = 0.095, P = 0.92. (ii) Prerequisite course grades: Grades from three required courses were collected. No significant differences were observed between cohorts Table 2. (iii) Pre-module written assessment: A 12-item assessment (10 multiple choice and two short answer questions) verified the retention of key concepts (central dogma, enzyme kinetics, and protein structure). The mean scores were comparable (control: 82.4%; experimental: 83.1%; P = 0.58). These three measures confirm equivalent academic standing and prior knowledge between cohorts at the start of the study.

TABLE 2.

Grades from three required courses between the 2022 and 2023 cohorts

Prerequisite course	Control group (mean % ± SD)	Experimental group (mean % ± SD)	P-value
Biochemistry	85.0 ± 5.8	85.2 ± 5.6	0.89
Molecular Biology	83.2 ± 6.4	83.7 ± 6.1	0.75
Microbiology	84.1 ± 6.2	84.5 ± 5.9	0.80

Ensuring fairness and minimizing bias

To ensure internal validity and fair comparison between groups, the following standardization measures were implemented (Table 3). Bias minimization strategies were as follows: (i) the same instructors taught both cohorts (instructor effect controlled), (ii) identical laboratory facilities and equipment were used across both years, (iii) double-blind grading of final reports and examinations, (iv) separate cohort years (2022 vs. 2023) prevented cross-contamination, and (v) standardized instructor scripts for core content delivery.

TABLE 3.

Standardization measures implemented in the work

Aspect	Control group (2022)	Experimental group (2023)	Standardization
Total instruction time	48 h	48 h	Identical
Laboratory access	Same facilities	Same facilities	Shared
Instructor team	Two lead instructors	Two lead instructors	Same personnel
Core learning materials	Basic protocols + textbook	Core protocols + readings	Equivalent
Common assessments	Lab performance rubric, final exam	Same rubrics	Identical
Grading	Double-blind	Double-blind	Same graders

Course implementation: week-by-week comparison

Weekly comparison of control and experimental group activities is shown in Table 4. To ensure fidelity to the PBL model, all instructors facilitating the 2023 cohort completed three pre-implementation workshops focusing on the following: (i) theoretical principles of PBL, (ii) facilitation strategies for open-ended inquiry (e.g., Socratic questioning and scaffolding), and (iii) techniques for managing group dynamics without providing direct solutions. The PBL module was implemented through a structured, multi-phase approach—Project Initiation Phase (Weeks 0–1): Students received the core project objective without detailed protocols. Instructor-led workshops introduced experimental design principles and literature search strategies. Teams submitted preliminary research proposals. Iterative Development Phase (Weeks 2–6): Mandatory interim team meetings scheduled at weeks 2, 4, and 6 (35 min each): team presentation (15 min covering milestones, data interpretation, challenges, and next steps); peer feedback (10 min); and instructor facilitation (10 min using Socratic questioning). Metacognitive Scaffolding: Students maintained individual reflection journals documenting problem-solving processes. When experimental failures occurred, instructors guided systematic troubleshooting using fishbone diagrams. Final Integration Phase (Weeks 7–8): Teams synthesized research processes in final reports, justifying deviations and design choices. Oral presentations included limitations analysis and future directions.

TABLE 4.

Comparison of control and experimental group activities

Weeks	Control group (traditional instruction)	Experimental group (PBL module)
0–2	Received step-by-step protocols for transformation and PCR verification, followed prescribed procedures	Conducted background research on ERG, formulated hypotheses, and performed transformation and PCR verification with self-designed approach
3–5	Followed fixed protocols for culture, harvest, lysis, protein analysis (SDS-PAGE), and ERG quantification	Selected independent variables; designed experimental procedures; performed culture, harvest, lysis, and analysis; and engaged in troubleshooting
6–7	Completed prescribed data collection and submitted standard lab report	Performed statistical analysis, prepared oral presentations, and submitted comprehensive research report
8	Final practical examination	Final practical examination and presentations

Assessment and evaluation instruments

Alignment of learning activities, outcomes, and assessments

Table 5 presents the alignment matrix demonstrating how learning activities support intended outcomes and their corresponding assessments.

TABLE 5.

Alignment matrix of learning activities, outcomes, and assessments

Learning activity	Intended learning outcome	Assessment method	Weight
Literature review and hypothesis generation	Formulate testable hypotheses	Research proposal	20%
Experimental design and variable selection	Design sound experimental strategy	Research proposal + lab performance	20% + 30%
Transformation, PCR, SDS-PAGE, HPLC	Demonstrate molecular biology proficiency	Lab performance (direct observation)	30%
Data collection and statistical analysis	Analyze and interpret experimental data	Final report	40%
Oral presentation	Communicate findings effectively	Oral presentation	10%
Troubleshooting and iterative refinement	Adapt based on data analysis	Reflection journals + final report	Embedded
Team collaboration	Work effectively in research teams	Peer evaluation + lab performance	Embedded

Development of assessment instruments

At the experiment’s end, all students responded to the questionnaires. A structured questionnaire was designed to measure student learning gains and satisfaction. Both cohorts (2022 and 2023) completed a questionnaire using a consistent five-level Likert scale (1 = strongly disagree to 5 = strongly agree). To assess and compare students’ learning gains in the course, in the study of the class of 2022 and 2023, we set a total of six questions: “Improvement in understanding ERG biosynthesis pathways; Enhanced ability to design and troubleshoot recombinant protein expression experiments; Increased proficiency in HPLC quantitative analysis; Better comprehension of metabolic engineering principles through practical application; Did the course improve your ability to connect experimental results with scientific literature? Has this experience increased your interest in pursuing research related to natural product biosynthesis?” These questions were on a five-point Likert-scale ranging from strongly disagree to strongly agree with five questions per attitude, sequentially assigned values of 1, 2, 3, 4, and 5.

To examine and compare the students’ satisfaction with the course, in the survey of the class of 2022 and 2023, we set a total of seven Likert scale questions with five levels, which were “Satisfaction with the ERG project design; Satisfaction with instructor guidance; Perceived relevance of the project to theoretical knowledge; Willingness to recommend this course to peers; Does the project-based approach enhance your laboratory engagement? Does the course meet your expectations for research training? Is the integration of molecular biology techniques in a single project more effective than traditional isolated experiments” The evaluated results of the four themes were collected to be examined: students’ gains in the course and their satisfaction with the course. Their reliability and validity were analyzed using SPSS 23.0 software.

Validity and reliability

All instruments underwent rigorous validation. Content validity was established through literature review and expert panel review (two biochemistry educators and one educational psychologist). Reliability was assessed using Cronbach’s alpha: the 2022 questionnaire achieved α = 0.81 (good internal consistency), while the expanded 2023 questionnaire reached α = 0.92 (excellent reliability). Both α values (0.81 and 0.92) exceed the conventional threshold for acceptable reliability (α ≥ 0.70), indicating that both versions of the questionnaire demonstrate good to excellent internal consistency. The increase in Cronbach’s alpha from 0.81 (2022 control) to 0.92 (2023 PBL) is likely due to the PBL model producing more consistent learning gains across students. PBL’s active, inquiry-driven approach reduces performance variability by engaging all students in hands-on problem-solving, leading to more homogeneous self-assessed competency levels and, consequently, higher internal consistency reliability.

Data collection and analysis

Academic performance data were collected through standardized rubric-based evaluations by the same instructor team across both years. Weighted components: laboratory performance (30%), research proposal (20%), oral presentation (10%), and final report (40%). Questionnaire data were collected anonymously during the final laboratory session (100% response rates). All data were double-entered into Excel and verified.

Descriptive statistics (means and standard deviations) were calculated for final grades and questionnaire items. Comparative analysis used independent samples t-tests, with assumptions of normality and homogeneity of variance verified. Statistical significance was set at P < 0.05, with effect sizes (Cohen’s d) calculated for significant comparisons. All analyses were performed using SPSS version 26.0.

Biosafety statement

The E. coli strain used in this study was E. coli BL21(DE3) and DH5α, non-pathogenic laboratory strain derived from E. coli K-12. This strain is classified as Biosafety Level 1 (BSL-1) according to the institutional and national guidelines. All experimental procedures were conducted in a standard teaching laboratory setting under BSL-1 containment. The following safety procedures were strictly followed: (i) work surfaces were disinfected with 75% ethanol before and after each experiment; (ii) all personnel wore disposable gloves and laboratory coats at all times; (iii) all materials that came into contact with E. coli cultures (including Petri dishes, pipette tips, and culture tubes) were autoclaved at 121°C for 20 min prior to disposal; and (iv) after each session, work areas were wiped with chlorine-containing disinfectant (500 ppm). These practices ensure safe handling and disposal of biological materials in an instructional setting.

RESULTS

Study design and baseline equivalence

A quasi-experimental design was employed to evaluate the effectiveness of the project-based learning (PBL) module against traditional laboratory instruction. The control cohort (Class of 2022, n = 30) followed a prescribed, step-by-step protocol for ergothioneine (ERG) production, while the experimental cohort (Class of 2023, n = 32) engaged in an open-ended PBL approach requiring experimental design and optimization (Fig. 3). To establish baseline equivalence, students’ prior cumulative Grade Point Averages (GPA) were compared between cohorts. An independent samples t-test revealed no statistically significant difference between the control group (M = 3.18, SD = 0.42) and the PBL group (M = 3.19, SD = 0.41); t (60) = 0.095, P = 0.92, confirming comparable academic standing at the start of the intervention.

Fig 3.

Study design and implementation framework. The diagram compares the traditional (2022) and PBL (2023) approaches across three phases: (i) project initiation and experimental design, (ii) implementation and data collection, and (iii) assessment and learning outcomes. Key pedagogical differences are highlighted, including student autonomy, experimental variables tested, and assessment components.

Comparative analysis of learning outcomes

Despite fundamental differences in pedagogical approach, the final course grades showed no significant difference between cohorts. The mean final grade for the control group was 85.0 ± 4.2, compared to 87.5 ± 3.5 for the PBL group (P = 0.31) (Fig. 4A). This equivalence in final grades is noteworthy because the assessment frameworks differed substantially between groups: the control group was evaluated on technical execution and report accuracy, while the PBL group was assessed across four research-aligned components (research proposal, laboratory performance, oral presentation, and final report). According to constructivist learning theory (20), learning is an active process where learners construct knowledge through experience and social interaction. The PBL module embodied this principle by engaging students in authentic research tasks without prescriptive protocols. Conversely, the traditional instruction cohort operated under a more objectivist paradigm, where knowledge is transmitted from the instructor to the learner through well-defined procedures. Despite these epistemological differences, both approaches produced statistically equivalent final grades. Biggs’ (21) theory of constructive alignment offers a compelling explanation for this finding. According to this framework, student learning outcomes are determined not by teaching method alone, but by the alignment between learning activities, assessment tasks, and intended learning outcomes. In the present study, both cohorts experienced internally consistent instructional systems: the control group’s step-by-step protocols aligned with its convergent assessments (technical accuracy and procedural compliance), while the PBL group’s open-ended inquiry aligned with its divergent assessments (research proposal, oral presentation, and comprehensive report). When alignment is achieved, different pedagogical approaches can produce comparable performance on their respective assessments.

Fig 4.

Academic performance comparison between control (2022) and PBL (2023) cohorts. (A) Final average grades showing no significant difference (P = 0.31). (B) Gender-disaggregated final grades illustrating the observed pattern reversal between cohorts.

Gender-disaggregated analysis revealed an interesting pattern: in the control cohort, female students outperformed male students (87.0 vs. 83.0), whereas in the PBL cohort, male students showed higher average grades than female students (89.0 vs. 86.5) (Fig. 4B). However, given the small sample size and lack of statistical power for subgroup analysis, this observation should be interpreted cautiously and warrants further investigation.

Student perceptions and learning achievement

Integrated analysis of student perceptions

To provide a coherent understanding of how the PBL module influenced student outcomes, we integrated the analysis of perceived learning gains, satisfaction, and engagement as interrelated dimensions of the student learning experience. All questionnaire items used five-level Likert scales unless otherwise noted. Overall, the PBL cohort consistently outperformed the control cohort across all measured dimensions (Fig. 5, detailed data in Tables S2 to S5). Mean score increases ranged from +0.13 to +0.45, with reliability analysis confirming internal consistency (Cronbach’s α: 2022 = 0.81, 2023 = 0.86) and validity metrics supporting construct validity (KMO: 2022 = 0.75, 2023 = 0.86; both P < 0.001). These psychometric properties confirm that observed differences reflect genuine variations in student perceptions rather than measurement artifacts.

Fig 5.

Comparative analysis of student-perceived learning gains between control (2022) and PBL (2023) cohorts. Data represent mean scores on seven-level Likert scales. The PBL cohort demonstrated consistently higher ratings across all dimensions, with the largest gains in research interest and conceptual understanding.

The two groups were compared in terms of affective outcomes and research interest

The most substantial gains were observed in affective and research-oriented outcomes. “Interest in pursuing research related to natural product biosynthesis” showed the largest increase (control: 3.86, PBL: 4.31, Δ = +0.45), indicating that the PBL approach effectively addressed a common limitation of traditional curricula, stimulating long-term research interest. This finding was reinforced by distribution analysis (Fig. 6A and B): the proportion of students rating their enthusiasm for scientific research as “very strong” increased from 58% to 66%, while lower-category ratings (I–III) consistently decreased, indicating a systematic positive shift.

Fig 6.

The evaluation of the program, including the result of the questionnaire. (A) Proportion of students from the class of 2022 and class of 2023 who believe the experimental design enhances enthusiasm for scientific research, rated from “I (very weak)” to “V (very strong).” (B) Proportion of students from the class of 2022 and class of 2023 who think the experimental design boosts professional competence, rated from “I (very weak) to V (very strong).”

Cognitive and skill-based outcomes of students in the two instructional groups were assessed

“Improvement in understanding ERG biosynthesis pathways” showed meaningful gains (control: 3.97, PBL: 4.22, Δ = +0.25), suggesting deeper conceptual understanding through hands-on inquiry. Skill-based dimensions demonstrated moderate but consistent improvements: experimental design ability (+0.13), HPLC proficiency (+0.13), and comprehension of metabolic engineering principles (+0.19). Notably, “Ability to connect experimental results with scientific literature” remained nearly equivalent between groups (Δ = −0.01), suggesting that this competency may be adequately addressed by both pedagogical approaches.

Students’ satisfaction and engagement were compared between the two instructional conditions

The PBL cohort also reported higher satisfaction across all evaluated dimensions. Mean satisfaction with experimental design increased from 4.26 to 4.50, instructor guidance from 4.20 to 4.44, willingness to recommend the course from 4.10 to 4.38, and perceived integration of molecular biology techniques from 4.18 to 4.44. The proportion of students rating professional competence as “very strong” rose from 56% to 62%, further supporting the positive shift in student perceptions.

Three key findings emerge from this integrated analysis. First, PBL produced consistently higher student perceptions across cognitive, affective, and satisfaction dimensions, with the largest effect on research interest, a construct closely linked to intrinsic motivation. Second, the parallel improvements in perceived learning gains and satisfaction suggest that these constructs are mutually reinforcing rather than independent. Third, the null finding for literature connection identifies a boundary condition where PBL does not confer additional advantage, indicating that different competencies may require different pedagogical approaches.

Experimental outcomes: pedagogical context and interpretation

The ERG production experiments yielded technically successful results in both cohorts, confirming the feasibility of the module for undergraduate laboratory instruction. However, the nature and interpretation of these experimental outcomes differed fundamentally between groups and must be understood within their pedagogical context.

Control cohort (2022): technical proficiency demonstration

For the control cohort, experimental procedures were standardized and prescribed. Students followed identical protocols for transformation, PCR verification, protein expression, and HPLC quantification. The experimental outcomes (detailed in Fig. S1) served primarily to demonstrate successful technical execution. All student groups successfully obtained positive transformants, confirmed by colony PCR, and detected target protein bands via SDS-PAGE. ERG production was confirmed in all groups, with yields ranging from approximately 150 to 300 mg/L.

These results validated that the experimental system was robust and reproducible under standardized conditions. The variation in yields, while present, was not systematically explored or optimized by students, as the prescribed protocol did not include experimental design components. Consequently, the experimental outcomes in the control cohort functioned as verification of technical competency rather than as products of scientific inquiry.

PBL cohort (2023): inquiry-driven experimental design

In contrast, the PBL cohort engaged with the experimental system as a platform for authentic scientific inquiry. Students were responsible for designing their own optimization strategies, selecting variables to test (IPTG concentration, pH, and temperature), and interpreting outcomes in the context of their hypotheses. This fundamental difference, prescribed execution versus inquiry-driven design, shaped both the process and the interpretation of experimental results.

The PBL cohort’s experimental outcomes (detailed in Fig. S2) revealed greater inter-group variability, wider ranges of ERG yields (from approximately 100 to 325 mg/L), and more diverse patterns of variable interactions. This variability was not a limitation but rather a pedagogical feature: it provided authentic opportunities for students to confront unexpected results, troubleshoot experimental failures, and refine their understanding of complex biological systems. The highest-yielding group achieved 325 ± 11 mg/L under self-optimized conditions, demonstrating that student-driven inquiry could match or exceed the outcomes of prescribed protocols while fostering deeper engagement with the scientific process

Distinguishing pedagogical impact from experimental artifact

A critical consideration in interpreting these findings is that the observed differences between cohorts are not primarily attributable to the experimental system itself, which remained largely consistent in terms of reagents, equipment, and technical procedures, but rather to how students engaged with that system. The ERG production platform served as a common context for both pedagogical approaches, but the nature of student interaction with this context differed fundamentally. For the control cohort, the experimental system was a demonstration tool—a means to verify that prescribed protocols yield expected results when correctly executed. For the PBL cohort, the same experimental system became an inquiry platform—a space for hypothesis generation, variable manipulation, data interpretation, and iterative refinement.

This distinction is crucial: the pedagogical intervention did not change the technical content of the experiments, but rather transformed students’ relationship to that content. The improved learning outcomes observed in the PBL cohort—greater research interest, deeper conceptual understanding, and enhanced troubleshooting skills—emerged not from different experimental results but from a fundamentally different process of engaging with those results.

The technical variability observed in the PBL cohort (e.g., wider yield ranges and more diverse optimization outcomes) reflects this pedagogical shift. In the control group, variability was minimized through standardization; in the PBL group, variability was embraced as a source of authentic learning. The fact that both cohorts achieved technically successful outcomes confirms that the PBL approach did not compromise technical skill acquisition—a common concern with open-ended pedagogies—while demonstrably enhancing higher-order cognitive and affective outcomes.

Synthesis and key findings

Three main findings emerge from this comparative analysis: First, the PBL approach enhanced higher-order learning outcomes without compromising technical competency. Despite fundamentally different pedagogical frameworks and assessment methods, final grades were equivalent between cohorts, and questionnaire data revealed consistent gains across conceptual, skill-based, and affective dimensions for the PBL group. Second, the most substantial impacts were observed in research-oriented and affective outcomes. The largest gains were in research interest (Δ = +0.45) and conceptual understanding (Δ = +0.25), suggesting that PBL’s primary value lies in transforming students’ relationship to scientific content—from passive recipients to active inquirers—rather than merely improving technical proficiency. Third, the experimental platform itself was a neutral context; the pedagogical approach determined learning outcomes. The same ERG production system, implemented with equivalent technical resources and faculty expertise, produced qualitatively different educational experiences depending on whether students engaged with it through prescribed protocols or inquiry-driven design. This finding underscores that pedagogical innovation, not experimental novelty, drove the observed improvements in learning outcomes.

DISCUSSION

Replacing “cookbook” labs with project-based learning: evidence from comparative outcomes

Traditional experimental courses often follow a “cookbook”-like model where students merely replicate predefined procedures (22–24). This approach, while ensuring experimental feasibility, may restrict students’ initiative and deeper thinking about experimental principles, condition optimization, and practical applications. Project-based learning (PBL) offers an alternative by centering on real-world challenges, fostering deeper understanding, critical thinking, and problem-solving skills (25–27).

Our results directly support this distinction. As reported in “Integrated analysis of student perceptions,” above, the PBL cohort consistently outperformed the control cohort across cognitive and affective dimensions. Specifically, “Improvement in understanding ERG biosynthesis pathways” showed a meaningful gain (control: 3.97, PBL: 4.22, Δ = +0.25), and “Better comprehension of metabolic engineering principles through practical application” increased from 3.90 to 4.09 (Δ = +0.19). These improvements were not observed in the control group, whose lower self-reported scores suggest that passive procedural execution, while sufficient for technical completion, does not equally promote conceptual integration.

Furthermore, the process-level evidence from “Experimental outcomes: pedagogical context and interpretation,” above, reinforces this interpretation. In the control cohort, experimental procedures were standardized and prescribed; students followed identical protocols without designing optimization strategies. In contrast, the PBL cohort engaged in inquiry-driven design, selecting variables such as IPTG concentration, pH, and temperature. This difference in process, not in technical content, likely explains the observed gains in conceptual understanding and research interest.

Utilizing experimental failures as teachable moments: evidence from troubleshooting skills and yield variability

A cornerstone of our pedagogical framework was the deliberate use of experimental setbacks as opportunities for cultivating scientific reasoning and resilience (28–30). Rather than providing immediate solutions, instructors facilitated structured, hypothesis-driven troubleshooting. Our data provide empirical support for this approach. As shown in “PBL cohort (2023): inquiry-driven experimental design,” above, the PBL cohort exhibited greater inter-group variability in ERG yields (ranging from approximately 100 to 325 mg/L) compared to the control cohort (150–300 mg/L). Critically, this variability was not treated as a limitation but as a pedagogical feature: it allowed students to confront unexpected results, troubleshoot failures, and refine their understanding. The highest-yielding PBL group achieved 325 ± 11 mg/L under self-optimized conditions, demonstrating that inquiry-driven design can match or exceed the outcomes of prescribed protocols.

Additionally, the questionnaire data (see “Integrated analysis of student perceptions,” above) directly support the effectiveness of this troubleshooting framework. “Enhanced ability to design and troubleshoot recombinant protein expression experiments” increased from 3.93 (control) to 4.06 (PBL, Δ = +0.13). While this gain is moderate, it is notable because troubleshooting skills are rarely assessed in traditional “cookbook” labs, where failures are minimized through standardization. The fact that PBL students reported higher troubleshooting confidence, despite encountering more variable experimental outcomes, suggests that exposure to authentic setbacks, when scaffolded appropriately, enhances rather than diminishes research competency.

A multidimensional assessment model for project-based learning: alignment and equivalence

Project-based teaching employs a more comprehensive assessment system (31–33) compared to traditional approaches that rely primarily on experimental reports or examination scores (33–35). Our assessment framework comprised oral presentation (10%), laboratory performance (30%), research proposal (20%), and final report (40%). Crucially, our results demonstrate that this multidimensional assessment did not disadvantage students relative to traditional assessment. As reported in “Comparative analysis of learning outcomes” (academic performance), above, final course grades showed no significant difference between cohorts (control: 85.0 ± 4.2, PBL: 87.5 ± 3.5, P = 0.31). This finding aligns with Biggs’ theory of constructive alignment (21): when learning activities, assessment tasks, and intended outcomes are internally consistent, different pedagogical approaches can produce comparable performance on their respective assessments. The control group’s step-by-step protocols aligned with convergent assessments (technical accuracy and procedural compliance), while the PBL group’s open-ended inquiry aligned with divergent assessments (proposal, presentation, and report). Thus, the equivalence in final grades should not be interpreted as equivalent learning outcomes; rather, it reflects that both instructional systems were internally coherent.

Quantitative evidence of the reformed curriculum’s impact: cognitive, skill-based, and affective outcomes

Our comparative analysis (see “Integrated analysis of student perceptions” and “Student perceptions and learning achievement,” above) provides concrete evidence for the multidimensional effectiveness of the teaching reform. As for cognitive outcomes, the PBL cohort outperformed the control group in “Improvement in understanding ERG biosynthesis pathways” (4.22 vs. 3.97) and “Comprehension of metabolic engineering principles” (4.09 vs. 3.90). These gaps highlight that the reformed curriculum moves beyond passive instruction by integrating contextualized practical tasks, helping students connect abstract concepts to hands-on operations.

As for the skill-based outcomes, the PBL cohort showed consistent advantages in laboratory skills (“Increased proficiency in HPLC quantitative analysis”: 4.19 vs. 4.06, Δ = +0.13) and research/investigation skills (“Enhanced ability to design and troubleshoot experiments”: 4.06 vs. 3.93, Δ = +0.13). These results indicate that the PBL approach shifts the curriculum from knowledge-centric to competency-driven. Notably, the smallest gap was observed in “Ability to connect experimental results with scientific literature” (4.12 vs. 4.13, Δ = −0.01). As reported in “Integrated analysis of student perceptions,” above, this near-equivalence suggests that literature connection skills may be adequately addressed by both pedagogical approaches. This finding identifies a boundary condition where PBL does not confer additional advantage, a specific direction for future refinement (e.g., integrating explicit training on literature analysis alongside inquiry-driven experiments).

Affective outcomes and research interest were also compared between the two instructional groups. The most pronounced difference emerged in “Interest in pursuing research related to natural product biosynthesis” (4.31 vs. 3.86, Δ = +0.45). As shown in Fig. 6A and B (see “Student perceptions and learning achievement,” above), the proportion of students rating their enthusiasm for scientific research as “very strong” increased from 58% (control) to 66% (PBL). This substantial gap highlights the reform’s unique role in transforming passive learning into active engagement, which is critical for inspiring students to pursue advanced studies or careers in STEM (36–39).

The experimental group’s questionnaire showed stronger reliability (Cronbach’s α: 0.86 vs. 0.81) and sampling adequacy (KMO: 0.86 vs. 0.75), confirming that observed differences reflect genuine variations in student perceptions rather than measurement artifact (see “Student perceptions and learning achievement,” above).

Distinguishing pedagogical impact from experimental artifact

A critical consideration in interpreting our findings is whether observed differences are attributable to the pedagogical intervention or to features of the experimental system itself. As explicitly addressed in “Distinguishing pedagogical impact from experimental artifact,” above, the ERG production platform was held constant across cohorts in terms of reagents, equipment, and technical procedures. The only systematic difference was how students engaged with that system: the control cohort used it as a demonstration tool (prescribed execution), while the PBL cohort used it as an inquiry platform (hypothesis generation, variable manipulation, and iterative refinement).

Three pieces of evidence support this interpretation. First, both cohorts achieved technically successful outcomes [see “Control cohort (2022): technical proficiency demonstration” and “PBL cohort (2023): inquiry-driven experimental design,” above], confirming that the PBL approach did not compromise technical skill acquisition. Second, the PBL cohort showed greater yield variability and more diverse optimization outcomes, features that emerged from student-driven design, not from differences in equipment or reagents. Third, the largest improvements in the PBL cohort were in higher-order outcomes (research interest Δ = +0.45, conceptual understanding Δ = +0.25) rather than in basic technical proficiency (HPLC proficiency Δ = +0.13). This pattern is consistent with a pedagogical effect rather than a mere artifact of the experimental platform.

Limitations and future directions

While the outcomes are promising, it is crucial to contextualize these findings within the study’s limitations. The implementation was conducted at one institution with specific resources and faculty expertise. Direct generalizability to other settings (e.g., community colleges, under-resourced universities) cannot be assumed. Additionally, the quasi-experimental design (Class of 2022 vs. Class of 2023) without randomization is a constraint. Potential inter-year variations (e.g., reagent batches and environmental conditions) were not statistically corrected for and should be considered when interpreting longitudinal comparisons. As noted in “Comparative analysis of learning outcomes,” above, the gender-disaggregated pattern (female students outperformed males in the control group, while male students showed higher average grades in the PBL group) should be interpreted cautiously due to the small sample size and lack of statistical power. This observation warrants further investigation with larger, balanced samples. The PBL module relied on access to molecular biology equipment and HPLC instrumentation, which may be a barrier for some institutions. Successful facilitation also requires instructors to shift from a directive to a guided-inquiry mindset, a transition that may require specific professional development that was available in our context. By acknowledging these boundaries, we aim to contribute not only a specific teaching model but also a framework for continued evaluation and collaborative improvement in the wider STEM education community.

Conclusion

In conclusion, preliminary results from the pedagogical reform integrating project-based learning with a comprehensive evaluation system in the ERG experimental course suggest its potential as an effective and promising educational approach. This reform not only significantly enhances students’' mastery of experimental techniques and theoretical knowledge related to ERG production, ranging from gene manipulation to fermentation optimization, but also comprehensively cultivates their scientific research capabilities. These include skills in experimental design, data analysis, and problem-solving. Moreover, it fosters essential soft skills such as teamwork spirit, effective communication, and scientific presentation abilities.

When compared with traditional teaching methods that often rely on rote procedure following and single outcome assessment, this reform better aligns with the core demands of modern higher education. It is dedicated to fostering innovative and practical talents who can not only grasp fundamental knowledge but also apply it creatively in real-world scientific research scenarios. By doing so, it not only enriches the learning experience of students in the field of biotechnology but also provides a valuable, replicable reference framework for the reform of similar experimental courses across various disciplines in higher education.

This study was conceived as a preliminary exploration to assess the immediate feasibility, student engagement, and perceived learning outcomes of the new module. While these initial metrics are encouraging, we concur that robust, objective evidence of learning, gauged through pre-/post-assessments and longitudinal tracking, is essential. Accordingly, we explicitly highlight the implementation of such comprehensive evaluation methods as a critical and necessary direction for future research.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jmbe.00109-26.

Text S1. jmbe.00109-26-s0001.docx.

Teaching syllabus.

Fig. S1. jmbe.00109-26-s0002.tif.

Work results for ERG production in E. coli and yield identification in class 2023.

Fig. S2. jmbe.00109-26-s0003.tif.

Work results for ERG production in E. coli and yield identification in class 2022.

Supplemental legends. jmbe.00109-26-s0004.docx.

Legends for Fig. S1 and S2.

Supplemental tables. jmbe.00109-26-s0005.docx.

Tables S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.