Xinysun Instrument (Shanghai) Co., Ltd.

Applications

Breadcrumb: Home > Applications > Lab Research

Colony Counting in Bacterial Competition Assays: How to Perform mCherry Fluorescence Sorting

Updated:2026-11-02Views:

Two strains are mixed 1:1 and co-cultured for six hours - which one comes out ahead? A CFU plate gives a total count but no ratio. A 2026 mLife paper on Acinetobacter baumannii put mCherry into the reference strain and read the plates on a LUYOR-3421 benchtop fluorescent protein excitation source, so that red and non-fluorescent colonies could be counted separately: a competitive index of 3.06 ± 0.88 at 3 hours and 3.45 ± 0.67 at 6 hours in BH broth. The article covers the experimental design, how the amplification strain was built, the competitive advantage seen across in vitro and mouse models, which module matches mCherry, and a six-step plating and reading procedure.

Two strains are mixed 1:1 and cultured together: which one comes out ahead after 6 hours? A CFU plate gives a total count but not a ratio. A 2026 paper in mLife on Acinetobacter baumannii put the red fluorescent protein mCherry into the reference strain and, after plating, read the colonies on a LUYOR-3421 benchtop fluorescent protein excitation source, so that red and non-fluorescent colonies could be counted separately. The methods section states it plainly: Colony ratios were determined using a LUYOR-3421 benchtop fluorescent protein excitation source and further confirmed by PCR with ISAba1-F/ISAba1-R primers. This article takes that step apart: how the label is set up, how the waveband is matched, how plates are read, and where the error comes from.

1. Why one strain in a competition experiment has to be marked

The logic of a competition assay is simple: mix the test strain and the reference strain in equal proportion, culture for a period, then see which way the ratio moves. A competitive index (CI) above 1 means the test strain is winning; below 1 means it is losing. The problem is that after mixing, the two strains grow identically on a plate, and the total CFU count cannot be split into two populations.

Three typing approaches are common, each with its own cost:

MethodApproachLimitation
Dual antibiotic markersEach strain carries one resistance and is plated on its own selective mediumClinical strains have complex resistance backgrounds, and two clean markers are often unavailable; two antibiotics also change growth rate
Colony-by-colony PCR20-24 single colonies picked at random per plate for colony PCR typingReliable, but laborious, and a low sampling fraction introduces statistical error
Fluorescent protein labelOne strain expresses mCherry stably; the whole plate is counted by colorNeeds an excitation source; reading relies on experience when expression is weak or colonies age

The advantage of a fluorescent label is that no colonies have to be picked, no second culture is needed, and the whole plate is split by eye - a plate can be counted in tens of seconds. The cost is that reading introduces new variables: source intensity, illumination angle, ambient light and colony age all shift the boundary between "shows red" and "does not show red". What decides data quality is therefore not whether a lamp is purchased but whether reading conditions are fixed and reproducible.

2. The experimental design: moving the amplification onto a plasmid

The work was done at the State Key Laboratory of Microbial Diversity and Innovative Utilization, Institute of Microbiology, Chinese Academy of Sciences, on Acinetobacter baumannii. The authors began by analyzing 7490 genomes (582 completed and 6908 draft), and found that a segment carrying crp and osmC flanked by two ISAba1 elements amplifies in multiple copies; this composite transposon was registered as Tn7691. Among the 582 completed genomes, amplification at this locus appeared in 19 strains (3.3%), with copy numbers of 2-8, and every strain carrying Tn7691 or showing amplification belonged to ST2 or its single-locus variants.

To reproduce the amplification in the laboratory, the authors cloned the ISAba1-crp-osmC sequence into the pKFAb plasmid and electroporated it into ATCC 19606 to obtain a mimic amplification strain; qPCR confirmed a 6.23 ± 1.97 fold rise in crp copy number. An empty-vector control and an mCherry-labelled control were built alongside:

Strain / plasmidFluorescent labelRole in the experiment
19606 (pKFAb)NoneEmpty-vector control for growth curve and transcriptome comparison
19606 (pKFAb:crp-osmC)NoneTest strain mimicking crp-osmC amplification
19606 (pKFAb:mCherry)Red fluorescenceReference strain for the competition assay, appearing red on plates
Clinical ST2 strain (pKF-ble / pKF-ble:crp-osmC)NoneBleomycin resistance in place of kanamycin; typed by colony PCR

Culture conditions were uniform: BHI broth, 37 °C, 200 rpm; plates were LB agar with kanamycin 50 mg/l (bleomycin 200 mg/l for the clinical strain work). The competition assay started at 5 × 106 CFU/ml, mixed 1:1, co-cultured at 37 °C and sampled at 0, 3 and 6 hours, then serially diluted and plated.

3. Results: how large the competitive advantage was

The conclusion comes before the method, because the reading procedure described later exists to serve these numbers:

SystemSampling timeCompetitive index (CI)
In vitro BHI co-culture3 h3.06 ± 0.88
In vitro BHI co-culture6 h3.45 ± 0.67
Mouse lung colonization24 hpi3.56 ± 0.97
Mouse intestinal colonization24 hpi2.12 ± 0.79
Mouse serum co-incubation3 h3.09 ± 1.84
Mouse serum co-incubation6 h9.00 ± 5.69
Lung in the cyclophosphamide immunosuppression model24 hpi5.54 ± 1.98

On the growth curve, the amplification strain began to separate at about 12 hours and reached a higher stationary-phase OD600; amplifying crp alone or osmC alone did not produce this effect, so the phenotype requires both genes in multiple copies. In the serum killing assay, survival of the amplification strain at 3 and 6 hours was 3.49 ± 1.34 and 7.05 ± 2.75 times that of the reference strain. The transcriptome pointed to three pathways: phenylalanine metabolism, siderophore synthesis and sulfur metabolism; at the same time biofilm formation fell and susceptibility to β-lactams, polymyxin B, ciprofloxacin and chloramphenicol rose, which the authors interpret as a trade-off between host fitness and resistance.

The magnitude of these numbers deserves attention: a CI between 2 and 9 means the ratio between the two populations differs by roughly two to nine fold, not a lopsided difference. Miscounting a dozen colonies on either side is enough to shift the CI by more than 10%. Stability of reading decides whether the conclusion holds.

4. Where 3421 sits in the workflow

The original text gives this step one sentence, but its position matters: it comes after dilution plating and before PCR confirmation, and its job is to split the colonies on a whole plate into "red" and "non-fluorescent" groups and yield a ratio; the subsequent PCR with ISAba1-F/ISAba1-R primers confirms that the red colonies do carry the corresponding plasmid. The in vivo work follows the same logic - lung tissue and fecal homogenates are plated, counted the same way by red fluorescence and non-fluorescence, and the competitive index calculated.

In other words, the instrument here performs reading, not imaging quantification. It does not need to produce presentable fluorescence photographs; it needs to let the operator separate two colony groups reliably under fixed conditions. That also explains the choice of a benchtop model over a handheld lamp: plates are read one after another, at time points across 0, 3 and 6 hours, and position, angle and brightness have to be identical every time.

5. Which module matches mCherry

mCherry has an excitation peak at 587 nm and an emission peak at 610 nm, which places it among the red fluorescent proteins. The GR module of LUYOR-3421 excites at 510-540 nm green and is paired with a 590 nm long-pass filter, which lands on the short-wave shoulder of the mCherry excitation curve.

Fluorescent protein / dyeExcitation peakEmission peak3421 moduleEmission filter
mCherry, DsRed, tdTomato587 nm610 nmGR (510-540 nm)590 nm long-pass
GFP, eGFP, FITC488 nm507 nmRB (440-460 nm)495 nm long-pass
YFP, Venus513 nm527 nmCY (490-515 nm)550 nm long-pass
CFP433 nm475 nmVI (400-415 nm)450 nm long-pass
DAPI, Hoechst358 nm461 nmUV (360-380 nm)415 nm long-pass

One technical detail needs stating: 510-540 nm is not the excitation peak of mCherry, and excitation there draws on the shoulder of the absorption curve, giving a weaker signal than 560-590 nm would. But mCherry is expressed strongly and packed densely inside a colony, so a yes/no classification is well within reach. For quantitative comparison of weakly expressing strains, or for measuring fluorescence intensity rather than counting, use longer-wavelength excitation or a fluorescence imaging system. That boundary should be kept clear, so a qualitative tool is not used as a quantitative instrument.

If an experiment carries both GFP and mCherry labels, fitting the RB and GR modules to one body is enough; they swap on BNC connectors in seconds, with no realignment of the light path.

Colony Counting in Bacterial Competition Assays: How to Perform mCherry Fluorescence Sorting(图1)

6. Six steps for reading plates

For the setting of this paper (90 mm plates, overnight culture, two colony groups mixed across the surface), the procedure can be organized into a reproducible sequence:

Step 1: control plate density

30-300 colonies per plate is a workable range. Too dense and colonies merge, so red and non-fluorescent cannot be counted separately; too sparse and the statistics become noisy. Make a 10-fold dilution series, plate two consecutive dilutions, and count the one that lands inside the range.

Step 2: fix the reading time

Count 18-24 hours after plating. Leave colonies longer and the red of mCherry dims as the colony ages, starting at the edges, which is how the same plate gets counted as 120 in the morning and 96 in the afternoon. If a re-check is needed, hold plates at 4 °C in the dark for no more than two days.

Step 3: start with the two control plates

The plate of the empty-vector strain (no fluorescence) sets the negative threshold - a signal that is invisible on this plate counts as non-fluorescent; the plate of the mCherry single-label strain sets the positive intensity. Both controls must be read at the same brightness setting, or the threshold means nothing.

Step 4: illuminate at an angle and avoid specular reflection

Bend the 500 mm gooseneck to 30-45° above the plate so the light spreads evenly over the whole dish. Direct frontal illumination produces a reflection band on the agar surface that hides part of the colonies; angled light also makes the fluorescence at colony edges stand out. Do not look into the lamp head while working.

Step 5: control ambient light rather than chasing total darkness

A darkroom is not needed. Switch off direct bench lighting, close the blinds and keep away from window reflections. The output of a 20 W LED is enough to make red colonies clearly distinguishable under ordinary room lighting, and holding ambient light at a consistent level is more practical than making the room fully dark each time.

Step 6: count by sector and photograph for the record

Draw a cross on the base of the dish to divide it into four sectors, count sector by sector and record. The 67 mm photography filter mounts directly on a DSLR, so the whole plate can be photographed - the image is not just an archive; if a result is disputed it can be re-read, which is far more reliable than recalling what was seen at the time.

Reading parameterSuggested settingReason
Colonies per plate30-300Avoids merging while keeping statistical power
Source angle30-45° obliqueReduces specular reflection from the agar surface
Brightness settingSame setting for the whole experimentKeeps thresholds comparable
Ambient lightDirect lamps off, away from windowsImproves red/non-red contrast
Reading time18-24 h after platingColonies mature, fluorescence stable
RecordingSector counts plus photographsResults can be re-checked and traced

7. Where counting error comes from

CI is obtained by dividing one colony count by another, so an error in either numerator or denominator is amplified directly. In practice the problems that keep appearing cluster in six places:

SymptomUsual causeRemedy
Red count too lowColonies aged and fluorescence decayed; dark red read as negativeCount within 24 hours; fix the brightness setting; re-read weak signals from photographs
Red count too highAgar or medium autofluorescence; negative threshold set too looselyUse the same medium batch; run an empty-vector plate with each batch to set the threshold
Both counts too lowPlasmid loss; antibiotic degradedMaintain selection with kanamycin 50 mg/l; prepare antibiotic fresh
Poor repeatabilityIllumination angle and distance changed between sessionsKeep the benchtop unit in a fixed position; mark the gooseneck angle; do not move the body
Colonies merge and cannot be separatedInsufficient dilution or uneven spreadingAdjust the dilution series; spread until the plate surface is dry
Results disagree with PCRFalse positives or false negatives in fluorescence readingPick colonies for colony PCR confirmation; the paper picked 20-24 per plate in the clinical strain group

One more point is easily overlooked: false negatives in fluorescence reading systematically push the test strain ratio down. If a red colony is recorded as negative because expression is weak, the CI moves in the opposite direction to the truth, and that shift does not cancel out at random. Control plates, a fixed brightness setting and photographic records are therefore not formalities; they are preconditions for the data being credible.

8. Why a benchtop model

LUYOR-3421 is positioned as a stereo microscope fluorescence adapter plus a standalone benchtop excitation source. In this kind of competition experiment the second form is what gets used: the body sits on the bench, the gooseneck brings the lamp head over the plate, and both hands are free for marking and counting.

ItemSpecification
Light source20 W high-power LED, large illuminated area, even spot
Wavelength modules6 groups: UV / VI / RB / RB-GO / CY / GR, BNC quick-change
Corresponding wavelengths365, 405, 450, 488, 505, 530 nm
Lamp arm500 mm flexible gooseneck, positionable at any angle
DimmingDimmer supplied as standard, continuously adjustable
Photography67 mm photography filter, mounts directly on a DSLR
Microscope fittingAdapter ring clamps to the stand, no modification, fast on and off
Power110-220 V AC, 50/60 Hz

Its second use comes in when detail matters: fit the adapter ring to a stereo microscope and insert the barrier filter below the objective, and expression distribution within a single colony can be examined under magnification, or the edge of a colony confirmed as genuinely red rather than confused by the agar background.

Other models in the series are positioned differently; select by scenario:

  • LUYOR-3415RG: dual-wavelength handheld, 0.9 kg, for screening while moving around an incubator room, a greenhouse or a field, one lamp per person.

  • LUYOR-3260RB: single-wavelength torch format, lighter and lower cost when the work involves GFP alone or RFP alone.

  • LUYOR-3124RB: large-area suspended format, for batch screening of seedling trays or large samples.

The criterion is really one question: does the sample come to the lamp, or does the lamp go to the sample? For plate counting, batch comparison and repeated reading across time points, use the benchtop unit; for work that moves between rooms and equipment, use the handheld.

9. Frequently asked questions

Does mCherry have to be excited at 587 nm?

Not necessarily. 587 nm is its excitation peak and gives the strongest signal, but colony reading needs just a yes/no decision, and 510-540 nm green is enough to excite a distinguishable red signal. What matters is using the same waveband and the same intensity each time, not chasing the peak.

Does plate reading have to be done in a darkroom?

No. Switching off direct lamps and staying away from window reflections is enough. What needs controlling is the stability of ambient light - a darkroom one day and a bright bench the next gives two different reading thresholds.

The red looks dim - does that mean there is no expression?

Rule out three things before anything else: whether the colonies are too old (red decays past 24 hours), whether the plate sat at 4 °C too long, and whether the plasmid was lost. If it is still dim after all three are excluded, expression itself may be low. In that case do not force a negative call; pick the colonies and run PCR instead.

Is PCR still needed after fluorescence counting?

Yes. This paper does exactly that: 3421 gives the colony ratio, and PCR with ISAba1-F/ISAba1-R confirms the identity of the red colonies; in the clinical strain group, 20-24 single colonies per plate were typed individually by PCR. Fluorescence reading provides efficiency, PCR provides identity.

How many colonies can one 90 mm plate carry?

30-300 is the usual range. Below 30 the statistics are noisy; above 300 colonies begin to merge and red signals overlap. In a mixed competition experiment the two colony groups are distributed at random, and the more evenly density is controlled, the better the repeatability of the CI.

3421 or 3415RG - how to choose?

By working style. If the reading bench is fixed, samples arrive in batches and photographs are needed for the record, choose 3421; if the work moves between incubator, greenhouse and field and one person carries the lamp while looking, choose 3415RG. Waveband coverage is the same; the difference is in form factor and illuminated area.

10. Related products and further reading

11. Closing remarks

The methodological value of this paper is not which lamp was used but that it writes down clearly a step that is easily left vague: competition assay typing rests on a fluorescent label, and reading that label rests on fixed conditions. 1:1 mixing, sampling at 3 and 6 hours, PCR confirmation of 20-24 colonies per plate - those details together are what make CI = 3.06 credible.

For laboratories running similar experiments, three lessons can be copied: the labelled strain needs stable expression, the reading bench needs to stay put, and every batch needs negative and positive control plates. Instrument selection is one link in that chain, but it decides whether what you count is data or noise. If you are setting up a reading workflow for bacterial competition or colonization work, send us the labelling strategy, plate format and batch size and we can look at waveband and model configuration together; you can also leave a message through contact us.

Related sections and models

This article sits in the laboratory research techniques section.

Other sections on the same product line are GFP fluorescence flashlights (models grouped by GFP/RFP/DsRed waveband, handheld and benchtop) and life science instruments (the entry point for fluorescence imaging, UV crosslinking, UV transmission and cell irradiation).

Models mentioned in this article: LUYOR-3421, LUYOR-3415RG, LUYOR-3260RB. Ranges, accuracy and accessory lists for each model are given on the product pages.

  • Official WeChat QR code

    Email: info@xinysun.com

  • Tel: 131-2093-6518

  • Address: 1525 Minqiang Road, Bldg 26, Songjiang, Shanghai

Official WeChat QR code Follow

Official WeChat QR code

Contact

131-2093-6518

Top