chromatography experiment class 11

If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

To log in and use all the features of Khan Academy, please enable JavaScript in your browser.

Class 11 Chemistry (India)

Course: class 11 chemistry (india)   >   unit 12.

• Simple and fractional distillations

Principles of chromatography

• Basics of chromatography
• Column chromatography
• Thin layer chromatography (TLC)
• Calculating retention factors for TLC
• Gas chromatography

• Take a few leaves and crush them in a mortar.
• Spot a drop of the leaf extract on a strip of chromatographic paper ~ 0.5 cm above the edge of the paper. Chromatographic paper is made of cellulose and is quite polar in nature.
• Place the strip of paper in a jar that contains a small volume of propanone (acetone). There should be just enough propanone that the edge of the paper dips in it comfortably. Place a lid on the jar to avoid any evaporation of the solvent.
• Let the solvent rise up the paper by capillary action. Remove the paper strip from the jar once the solvent has reached the ‘solvent front’ level. 5) What do you think you will notice?
• Higher the adsorption to the stationary phase, the slower the molecule will move through the column.
• Higher the solubility in the mobile phase, the faster the molecule will move through the column.

Different types of chromatography

Thin layer chromatography (TLC): Retention factors (R f ‍   )

• The component that travels the least distance on the TLC plate is the most polar, since it binds to the silica most tightly.
• The component that travels the maximum distance is the least polar; it binds to the silica least tightly and is most soluble in the non-polar solvent (mobile phase), and hence moves up the plate with the solvent.

Attribution:

Want to join the conversation.

• Upvote Button navigates to signup page
• Downvote Button navigates to signup page
• Flag Button navigates to signup page

• Science Notes Posts
• Contact Science Notes
• Todd Helmenstine Biography
• Anne Helmenstine Biography
• Free Printable Periodic Tables (PDF and PNG)
• Periodic Table Wallpapers
• Interactive Periodic Table
• Periodic Table Posters
• Science Experiments for Kids
• How to Grow Crystals
• Chemistry Projects
• Fire and Flames Projects
• Holiday Science
• Chemistry Problems With Answers
• Physics Problems
• Unit Conversion Example Problems
• Chemistry Worksheets
• Biology Worksheets
• Periodic Table Worksheets
• Physical Science Worksheets
• Science Lab Worksheets
• My Amazon Books

Leaf Chromatography Experiment – Easy Paper Chromatography

Leaf chromatography is paper chromatography using leaves. Paper chromatography is a separation technique. When applied to leaves, it separates the pigment molecules mostly according to their size. The main pigment molecule in green leaves is chlorophyll, which performs photosynthesis in the plant. Other pigments also occur, such as carotenoids and anthocyanins. When leaves change color in the fall , the amount and type of pigment molecules changes. Leaf chromatography is a fun science project that lets you see these different pigments.

Leaf Chromatography Materials

You only need a few simple materials for the leaf chromatography project:

• Rubbing alcohol (isopropyl alcohol)
• Coffee filters or thick paper towels
• Small clear jars or glasses with lids (or plastic wrap to cover the jars)
• Shallow pan
• Kitchen utensils

You can use any leaves for this project. A single plant leaf contains several pigment molecules, but for the most colors, use a variety of leaves. Or, collect several of each kind of leaf and compare them to each other. Good choices are colorful autumn leaves or chopped spinach.

Perform Paper Chromatography on Leaves

The key steps are breaking open the cells in leaves and extracting the pigment molecule and then separating the pigment using the alcohol and paper.

• Finely chop 2-3 leaves or several small leaves. If available, use a blender to break open the plant cells. The pigment molecules are in the chloroplasts of the cells, which are organelles encased within the plant cell walls. The more you break up the leave, the more pigment you’ll collect.
• Add enough alcohol to just cover the leaves.
• If you have more samples of leaves, repeat this process.
• Cover the container of leaves and alcohol and set it in a shallow pan filled with enough hot tap water to surround and heat the container. You don’t want water getting into your container of leaves.
• Replace the hot water with fresh water as it cools. Swirl the container of leaves around from time to time to aid the pigment extraction into the alcohol. The extraction is ready when the alcohol is deeply colored. The darker its color, the brighter the resulting chromatogram.
• Cut a long strip of coffee filter or sturdy paper towel for each chromatography jar. Paper with an open mesh (like a paper towel) works quickly, but paper with a denser mesh (like a coffee filter) is slower but gives a better pigment separation.
• Place a strip of paper into jar, with one end in the leaf and alcohol mixture and the other end extending upward and out of the jar.
• The alcohol moves via capillary action and evaporation, pulling the pigment molecules along with it. Ultimately, you get bands of color, each containing different pigments. After 30 to 90 minutes (or whenever you achieve pigment separation), remove the paper strips and let them dry.

How Leaf Chromatography Works

Paper chromatography separates pigments in leaf cells on the basis of three criteria:

• Molecule size

Solubility is a measure of how well a pigment molecule dissolves in the sol vent. In this project, the solvent is alcohol . Crushing the leaves breaks open cells so pigments interact with alcohol. Only molecules that are soluble in alcohol migrate with it up the paper.

Assuming a pigment is soluble, the biggest factor in how far it travels up the paper is particle size. Smaller molecules travel further up the paper than larger molecules. Small molecules fit between fibers in the paper more easily than big ones. So, they take a more direct path through the paper and get further in less time. Large molecules slowly work their way through the paper. In the beginning, not much space separates large and small molecules. The paper needs to be long enough that the different-sized molecules have enough time to separate enough to tell them apart.

Paper consists of cellulose, a polysaccharide found in wood, cotton, and other plants. Cellulose is a polar molecule . Polar molecules stick to cellulose and don’t travel very far in paper chromatography. Nonpolar molecules aren’t attracted to cellulose, so they travel further.

Of course, none of this matters if the solvent doesn’t move through the paper. Alcohol moves through paper via capillary action . The adhesive force between the liquid and the paper is greater than the cohesive force of the solvent molecules. So, the alcohol moves, carrying more alcohol and the pigment molecules along with it.

Interpreting the Chromatogram

• The smallest pigment molecules are the ones that traveled the greatest distance. The largest molecules are the ones that traveled the least distance.
• If you compare chromatograms from different jars, you can identify common pigments in their leaves. All things being equal, the lines made by the pigments should be the same distance from the origin as each other. But, usually conditions are not exactly the same, so you compare colors of lines and whether they traveled a short or long distance.
• Try identifying the pigments responsible for the colors.

There are three broad classes of plant pigments: porphyrins, carotenoids, and flavonoids. The main porphyrins are chlorophyll molecules. There are actually multiple forms of chlorophyll, but you can recognize them because they are green. Carotenoids include carotene (yellow or orange), lycopene (orange or red), and xanthophyll (yellow). Flavonoids include flavone and flavonol (both yellow) and anthocyanin (red, purple, or even blue).

Experiment Ideas

• Collect leaves from a single tree or species of tree as they change color in the fall. Compare chromatograms from different colors of leaves. Are the same pigments always present in the leaves? Some plants produce the same pigments, just in differing amounts. Other plants start producing different pigments as the seasons change.
• Compare the pigments in leaves of different kinds of trees.
• Separate leaves according to color and perform leaf chromatography on the different sets. See if you can tell the color of leaves just by looking at the relative amount of different pigments.
• The solvent you use affects the pigments you see. Repeat the experiment using acetone (nail polish remover) instead of alcohol.
• Block, Richard J.; Durrum, Emmett L.; Zweig, Gunter (1955).  A Manual of Paper Chromatography and Paper Electrophoresis . Elsevier. ISBN 978-1-4832-7680-9.
• Ettre, L.S.; Zlatkis, A. (eds.) (2011). 75 Years of Chromatography: A Historical Dialogue . Elsevier. ISBN 978-0-08-085817-3.
• Gross, J. (1991). Pigments in Vegetables: Chlorophylls and Carotenoids . Van Nostrand Reinhold. ISBN 978-0442006570.
• Haslam, Edwin (2007). “Vegetable tannins – Lessons of a phytochemical lifetime.”  Phytochemistry . 68 (22–24): 2713–21. doi: 10.1016/j.phytochem.2007.09.009
• McMurry, J. (2011). Organic chemistry With Biological Applications (2nd ed.). Belmont, CA: Brooks/Cole. ISBN 9780495391470.

Related Posts

Talk to our experts

1800-120-456-456

Separation of Plant Pigments by Paper Chromatography

Plant pigments – an introduction.

A compound that absorbs light is called a pigment. Chlorophylls a and b are primary photosynthetic pigments that absorb light for photosynthesis . The accessory pigments carotenoids and xanthophyll absorb light and pass it to chlorophyll a. Even though chlorophyll is the primary pigment , the other pigments are essential to the plant's ability to produce colour and engage in photosynthesis because they absorb each light differently and effectively across the electromagnetic spectrum.

Paper Chromatography

Chromatography, which means "colour writing," is a Greek term that is formed from the words "chromo" and "graph". Chromatography enables the separation of the constituent parts of a given mixture, enabling scientists to observe and produce findings and theories. 

Paper chromatography is a method for classifying dissolved substances according to how soluble they are in a given solvent, such as chlorophyll, carotene, and xanthophyll. Paper chromatography can be used to separate the colours in plant cells. The stationary element in chromatography paper permits the reaction between the solute and solvent to take place and produce results.

Leaf Chromatography

The separation of leaf colours using chromatography is known as leaf chromatography. Leaf chromatography is an experiment that is conducted to determine the colour of the photosynthetic pigments.

The experiment is conducted to learn about the pigments in the leaf, and it is mostly done by using paper and thin-layer chromatography. Let’s discuss some brief points of leaf chromatography.

There is a procedure by which this experiment is conducted in labs.

Sample leaves should be crushed into small pieces and put in a mortar for pestle grinding. Add solvent and keep using the pestle to crush.

Then, carefully draw a pencil line 1 cm from the bottom of the chromatography paper, spot a little amount of leaf extract repeatedly onto the centre of the line, and let each spot dry. 

Make sure the paper dips into the solvent but the spot of leaf extract doesn't by suspending it using a pin attached to a bung within a test tube with a 1 cm depth of solvent. 

The solvent is allowed to run up the paper until it is close to the bung, at which point the paper is removed. The solvent's location is marked, and the paper is allowed to dry. 

The final chromatography paper is known as a chromatogram, and it may be photographed to determine the exact position of each pigment. Next, determine the Rf value for each pigment spot on the chromatogram.

The retention factor is pronounced Rf. The retention factor is calculated by dividing the component's travel distance by the solvent's travel distance.

The colour dissolves as the alcohol goes through the filter paper. Some pigments in the leaf travel more quickly than others because of their properties. 

The pigment's movement rate is measured by the Rf (retention factor) value. Rf value = distance transported by pigment from origin to centre of pigment spot/distance from the origin to the solvent front. By applying this formula, you can determine the Rf value.

The pigments in the plant's leaf are separated by paper chromatography, i.e., separation chromatography. It is the same as a leaf chromatography experiment. The process of paper chromatography is also the same as the leaf chromatography experiment. 

Separation of Chlorophyll Pigments by Paper Chromatography

The chlorophyll molecule is present in the leaf and can be separated by using paper chromatography. The paper chromatography separates the pigments in the leaf based on the distance travelled by pigment molecules on the paper in a nonpolar solvent.

Separation of Plant Pigments by Paper Chromatography Diagram

The Experimental Setup of Paper Chromatography

Chromatogram Report

The final chromatography paper is known as a chromatogram, and it may be photographed to determine the exact position of each pigment. The pattern of pigment spots on the chromatography paper at the conclusion of the experiment is called a chromatogram. Along with the alcohol, the pigments also migrate along the strips of paper.

Chromatography Conclusion

Carotene is identified as having the lowest molecular weight by its yellow to orange tint near the top of the paper. In the pigment separation of chlorophyll, chlorophyll may be distinguished by its blue or dark green hue. When chlorophyll pigments are separated, the colour yellow-light green identifies chlorophyll B. In the chromatography solvent, xanthophyll is more soluble since it has gone up the paper. This describes the conclusion of paper chromatography.

Conclusion 

The pigments are light-absorbing molecules and are separated by using paper chromatography techniques in the lab. The pigments move on the paper based on their solubility in the solvent. Along with the alcohol, pigments also migrate along the strips of paper. Some pigments in the leaf travel more quickly than others because of their properties.

FAQs on Separation of Plant Pigments by Paper Chromatography

1. What is the spectrum of absorption and action?

Infrared and visible electromagnetic energy is absorbed by photochromic pigments. The term "absorption spectrum" refers to a graph that displays the degree of light absorption by various pigments in sunlight as a function of wavelength. These pigments in leaves absorb various light wavelengths. Chlorophyll is, especially, sensitive to the colours red and blue. The action spectrum is a graph that demonstrates the degree to which various light wavelengths may effectively catalyse a photochemical process. Green has the greatest detrimental impact on light absorption. The selectively permeable membrane does not absorb green, so the chloroplast reflects it.

2. What is chromatography and its different types?

Different molecules can be distinguished from one another using the process of chromatography. It enables the separation of a mixture into its constituent parts when it is passed through a stationary phase by a mobile phase. The phase that is fixed in place in a chromatography column or plate is known as the stationary phase. The analysed mixture is transported through the stationary phase by the mobile phase. These characteristics—shape, size, charge, mass, adsorption, and solubility—are the basis for this distinction. 

Column chromatography, paper chromatography, partition chromatography, and thin-layer chromatography are some examples of chromatography.

3. What role does pigment play in photosynthesis?

Molecules called photosynthetic pigments absorb some wavelengths of light while reflecting others. We perceive them to be the colour of the wavelength that they reflect. To absorb the lightest energy possible, a variety of pigments work together. They are kept in thylakoid membranes, arrayed in photosystems, which are funnel-shaped structures maintained in place by proteins. During photosynthesis, the free electrons in the pigments' chemical structures transform into high-energy electrons, releasing the energy that they had previously received from light and transferring it to other molecules.

• Skip to primary navigation
• Skip to main content
• Skip to primary sidebar

• FREE Experiments
• Kitchen Science
• Climate Change
• Egg Experiments
• Fairy Tale Science
• Edible Science
• Human Health
• Inspirational Women
• Forces and Motion
• Science Fair Projects
• STEM Challenges
• Science Sparks Books
• Contact Science Sparks
• Science Resources for Home and School

Paper Chromatography Experiment

March 17, 2021 By Emma Vanstone Leave a Comment

This simple felt tip pen paper chromatography experiment is a great way to learn about this particular method of separating mixtures .

WHAT IS CHROMATOGRAPHY?

Chromatography   is a technique used to separate mixtures. Information from a chromatography investigation can also be used to identify different substances.

In chromatography, the mixture is passed through another substance, in this case, filter paper. The different-coloured ink particles travel at different speeds through the filter paper, allowing the constituent colours of the pen ink to be seen.

All types of chromatography have two phases: a mobile phase where the molecules can move and a stationary phase where they can’t move. In the case of paper chromatography, the stationary phase is the filter paper, and the mobile phase is the solvent ( water ).

The more soluble the ink molecules, the further they are carried up the paper.

The video below shows chromatography in action.

You’ll need:

Filter paper or paper towel

Felt tip pens – not washable or permanent

A container – glass, jar or plate

Instructions

Pour a small amount of water onto a plate or into the bottom of a jar.

Find a way to suspend the filter paper over the water so that just the very bottom touches the water. If you do the experiment in a jar, the easiest way to do this is to wrap the top of the filter paper around a pencil, clip it in place, and suspend it over the top of the jar.

Our LEGO holder worked well, too!

Use the felt tip pens to draw a small circle about 1cm from the bottom of the filter paper with each colour pen you want to test.

Suspend the filter paper in the water and watch as the ink moves up the filter paper.

You should end up with something like this! The end result is called a chromatogram.

What happens if you use washable pens?

If the inks are washable, they tend to contain just one type of ink, so there is no separation of colour.

Below, only a couple of the inks have separated compared to the non-washable pens above.

Why does chromatography work?

When the filter paper containing the ink spots is placed in the solvent ( in this case, water ), the dyes travel through the paper.

Different dyes in ink travel through the chromatography filter paper at different speeds. The most soluble colours dissolve and travel further and faster than less soluble dyes, which stick to the paper more.

I’ve created a free instruction sheet and chromatography experiment write up to make the activity even easier.

Extension task

Experiment with different types and colours of pens. Depending on the type of ink used, some will work better than others.

Try chromatography with sweets .

Steamstational also has a great leaf chromatography investigation.

More separation experiments

Clean up water by making your own filter .

Separate water and sand by evaporation .

Make colourful salt crystals by separating salt and water.

Separate liquid mixtures with a bicycle centrifuge .

Last Updated on May 20, 2024 by Emma Vanstone

Safety Notice

Science Sparks ( Wild Sparks Enterprises Ltd ) are not liable for the actions of activity of any person who uses the information in this resource or in any of the suggested further resources. Science Sparks assume no liability with regard to injuries or damage to property that may occur as a result of using the information and carrying out the practical activities contained in this resource or in any of the suggested further resources.

These activities are designed to be carried out by children working with a parent, guardian or other appropriate adult. The adult involved is fully responsible for ensuring that the activities are carried out safely.

Reader Interactions

Leave a reply cancel reply.

Your email address will not be published. Required fields are marked *

• Class 6 Maths
• Class 6 Science
• Class 6 Social Science
• Class 6 English
• Class 7 Maths
• Class 7 Science
• Class 7 Social Science
• Class 7 English
• Class 8 Maths
• Class 8 Science
• Class 8 Social Science
• Class 8 English
• Class 9 Maths
• Class 9 Science
• Class 9 Social Science
• Class 9 English
• Class 10 Maths
• Class 10 Science
• Class 10 Social Science
• Class 10 English
• Class 11 Maths
• Class 11 Computer Science (Python)
• Class 11 English
• Class 12 Maths
• Class 12 English
• Class 12 Economics
• Class 12 Accountancy
• Class 12 Physics
• Class 12 Chemistry
• Class 12 Biology
• Class 12 Computer Science (Python)
• Class 12 Physical Education
• GST and Accounting Course
• Excel Course
• Tally Course
• Finance and CMA Data Course
• Payroll Course

Interesting

• Learn English
• Learn Excel
• Learn Tally
• Learn GST (Goods and Services Tax)
• Learn Accounting and Finance
• GST Tax Invoice Format
• Accounts Tax Practical
• Tally Ledger List
• GSTR 2A - JSON to Excel

Are you in school ? Do you love Teachoo?

We would love to talk to you! Please fill this form so that we can contact you

• NCERT Questions
• Questions from Inside the chapter
• Examples from NCERT Book
• Teachoo Questions

Chromatography

Last updated at April 16, 2024 by Teachoo

Chromatography is used to separate 2 or more dissolved solids in a solution, but these dissolved solids are in very small quantities . Example - Different Dyes from Blank Ink

What is chromatography?

It is a technique used to separate different solutes that dissolve in the same solvent.

This technique is used for separation of different colors .

• It is used to separate different colors of dye.
• It is used to separate drugs from blood
• It is used to separate different colours from pigments

Central Idea behind Chromatography

Different solutes have different solubilities in a solvent. And these solutes are separated using Chromatography

Let's look at a Case where Chromatography is used

Separating Different Colors of Dye

Ink has water as solvent and dye is soluble(solute) in it.

Is DYE  made of Single color?

• Dye in ink is made up of different colors mixed together.

How to Separate Different Colors of Dye

• Take a small strip of filter paper
• Draw a line on it (around 3 cm from bottom)
• Put a small drop of ink at the center of the line using a sketch pen.
• Now take  a beaker filled with some water
• Lower the filter paper in the beaker
• Lower end of filter paper should dip in water but ink spot should be just above the water
• After some times, we see water rise up over the filter paper
• Different colors of ink are also visible at different heights

When water rises up on filter paper, it takes the dye of ink along with it.

• The component which is more soluble in water rises faster and is seen at higher position
• The component which is less soluble is seen below at lower position

Maninder Singh

CA Maninder Singh is a Chartered Accountant for the past 14 years and a teacher from the past 18 years. He teaches Science, Economics, Accounting and English at Teachoo

Hi, it looks like you're using AdBlock :(

Please login to view more pages. it's free :), solve all your doubts with teachoo black.

Our objective is to separate and study plant pigments by paper chromatography.

Before going into detail, let’s understand the role of pigments in plants.

Photosynthetic plants convert light energy from the sun to chemical food energy.  During photosynthesis, molecules referred to as pigments are used to capture light energy.  Pigments are chemical compounds which reflect only certain wavelengths of visible light.  Plant leaves contain four primary pigments: chlorophyll a (dark green), chlorophyll b (yellowish-green), xanthophylls (yellow) and carotenoids (orange). 

To separate and visualize the four primary pigments of green plants, we can use a simple technique called chromatography.

What is Chromatography?

Chromatography is a technique used to separate molecules on the basis of differences in size, shape, mass, charge, solubility and adsorption properties. The term chromatography is derived from Greek words Chroma-colour and Graphe-write. There are many types of chromatography: paper chromatography, column chromatography, thin layer chromatography and partition chromatography. These techniques involve the interaction between three components: the mixture to be separated, a solid phase and a solvent.

How does paper chromatography work?

In paper chromatography, the mixture is spotted onto the paper, dried and the solvent is allowed to flow along the sheet by capillary attraction.  As the solvent slowly moves through the paper, the different compounds of the mixture separate into different coloured spots. The paper is dried and the position of different compounds is visualized. The principle behind the paper chromatography is that the most soluble substances move further on the filter paper than the least soluble substances. Different plant pigments can be separated by using the technique of paper chromatography.

What is Retention Factor or Rf value?

Retention factor or R_f value is applied in chromatography to make the technique more scientific than a mere analysis. The retention factor or Rf is defined as the distance travelled by the compound divided by the distance traveled by the solvent. R_f=(Distance travelled by the compound)/(Distance travelled by the solvent)

Diagrammatic example that demonstrates Rf value:

Learning Outcomes

• Students will understand the principle behind chromatography techniques.
• Students will learn about different types of pigments occurring in a plant leaves.
• Students will learn how to calculate the Retention factor.
• Students will be able to do the experiment more accurately in the real lab once they understand the steps through the animation and simulation.

Developed by Amrita Vishwa Vidyapeetham & CDAC Mumbai. Funded by MeitY (Ministry of Electronics & Information Technology)

English हिंदी മലയാളം मराठी

Notes For Free

One Stop Repository

Chromatography

AIM : To separate the coloured components present in a mixture of red and green ink by ascending paper chromatography and find their R f  values.

In this type of chromatography, a special adsorbent paper (Whatman filter paper) is used. Moisture adsorbed bon this Whatman filter paper acts as the stationary phase and the solvent acts as the   mobile phase .  The mixture to be separated is spotted at one end of the paper. This paper is then developed in a particular solvent by placing the paper in a gas jar, taking care that the spot is above the solvent. The solvent rises due to capillary action and the components get separated out as they rise up with the solvent at different rates. The developed paper is called a chromatogram.

R f  (retention factor) values are then calculated, which is the ratio of the distance moved by the component to the distance moved by the solvent front.

R f  =  Distance traveled by the component          Distance traveled by the solvent front

OBSERVATIONS AND CALCULATIONS: ( ON THE BLANK PAGE, USING A PENCIL)

RESULT:  (ON RULED SIDE )  –

The components of the mixture (red and green colour) separate in the form of spots lying between the origin line and solvent front.

R f(green)  = d g /d s = 3.8/4 =0.95

R f(red)  = d r /d s = 2/4= 0.5

Precautions

• The strip should not touch the walls of the jar.
• Do not disturb the jar after putting the strip in it.
• Allow the colours to seperate before pulling the strip out.

You can also get Class XII Practicals on  Biology ,  Physics , and  Physical Education .

12 Replies to “Chromatography”

Yo dawg this shit is real cool kuku

Awesome Really helpful and saved my marks.

It was really very helpful for me! Please post the experiments of Salt Analysis of class XI. ✨

Really helpful…🙏

Can you give this same for orange ink. Plss….

Thankyou brother 😁

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Med Sci Monit
• PMC11302215

Gas Chromatography-Ion Mobility Spectrometry Reveals Acetoin as a Biomarker for Carbapenemase-Producing Klebsiella pneumoniae

1 Jiangxi Province Key Laboratory of Immunology and Inflammation, Jiangxi Provincial Clinical Research Center for Laboratory Medicine, Department of Clinical Laboratory, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, Jiangxi, PR China

Yunwei Zheng

Chuwen zhao.

2 School of Public Health, Nanchang University, Nanchang, Jiangxi, PR China

Yaping Hang

Youling fang.

This study aimed to detect the volatile organic compound (VOC), 3-hydroxy-2-butanone (acetoin) using gas chromatography-ion mobility spectrometry (GC-IMS) in antimicrobial-resistant Klebsiella pneumoniae ( K. pneumoniae ) carbapenemase (KPC)-producing bacteria.

Material/Methods

Using stromal fluid of blood culture bottles (BacT/ALERT ® SA) as the medium, 3-hydroxy-2-butanone (acetoin) released by K. pneumoniae during growth was detected using GC-IMS. The impact of imipenem (IPM) and carbapenemase inhibitors [avibactam sodium or pyridine-2,6-dicarboxylic acid (DPA)] on the emission of 3-hydroxy-2-butanone (acetoin) from various carbapenemase-producing K. pneumoniae was further investigated. Subsequently, VOCal software was used to generate a pseudo-3D plot of 3-hydroxy-2-butanone (acetoin), and the relative peak volumes were exported for data analysis. Standard strains served as references, and the findings were validated with clinical isolates.

The pattern of temporal changes in the 3-hydroxy-2-butanone (acetoin) release from K. pneumoniae in the absence of IPM was consistent with the growth curve. After the IPM addition, carbapenemase-positive strains released significantly higher contents of 3-hydroxy-2-butanone (acetoin) than carbapenemase-negative strains at the late exponential growth phase (T2). Notably, adding avibactam sodium significantly decreased the 3-hydroxy-2-butanone (acetoin) content released from the class A carbapenemase-producing strains as compared to the absence of the carbapenemase inhibitor. Conversely, adding DPA significantly decreased the 3-hydroxy-2-butanone (acetoin) content released from the class B carbapenemase-producing strains (both standard and clinical strains, all P <0.05).

Conclusions

This study demonstrated the potential of 3-hydroxy-2-butanone (acetoin) as a VOC biomarker for detecting carbapenemase-producing K. pneumoniae , as revealed by GC-IMS analysis.

Introduction

Recently, the extensive use of carbapenem antibiotics in clinical practice has led to a concerning rise in carbapenem resistance among Klebsiella pneumoniae ( K. pneumoniae ). Since the 1990s, the prevalence of carbapenem-resistant K. pneumoniae (CRKP) has progressively increased globally, resulting in high morbidity and mortality rates associated with these infections [ 1 ]. Furthermore, the emergence of antimicrobial-resistant K. pneumoniae carbapenemase (KPC)-producing bacteria has become a pressing global issue [ 2 – 5 ].

Carbapenemase production is the main mechanism of resistance to carbapenem antibiotics in CRKP strains, with K. pneumoniae carbapenemase (KPC-type), New Delhi metallo-beta-lactamase (NDM-type) carbapenemase, and oxacillinase-48 (OXA-48-type) carbapenemase being the most prevalent types [ 6 , 7 ]. In China, the main genetic determinant of CRKP is K. pneumoniae carbapenemase-2 (KPC-2), accounting for approximately 70% of cases [ 8 ]. Carbapenemases are classified according to the Ambler molecular classification into classes A, B, and D. Class A and D carbapenemases are serine-like enzymes, and class B are metalloenzymes [ 9 ]. Notably, antimicrobial drugs exhibited varying in vitro antimicrobial activities against different carbapenemase-producing strains [ 10 ]. Metalloenzyme-producing CRKP strains often show sensitivity to aztreonam. Conversely, the broad-spectrum antibiotic ceftazidime-avibactam (CAZ-AVI) possesses robust antimicrobial properties against carbapenemase (serine-like enzymes)-producing CRKP strains but lacks efficacy against metalloenzyme-producing CRKP counterparts. Therefore, the accurate and prompt identification and classification of carbapenemases in CRKP are crucial for the appropriate clinical administration of anti-infective medication.

Currently, the laboratory tests for carbapenemases rely on the modified carbapenem inactivation method (mCIM) and the Carbapenemase Non-Phenotypic (Carba NP) test, as recommended by the Clinical and Laboratory Standards Institute (CLSI) [ 11 , 12 ]. However, these techniques require routine cultivation of pure colonies, with certain methods also necessitating overnight culture for a minimum of 1-2 days [ 13 ]. This delay hinders prompt clinical diagnosis and treatment. Despite developing various carbapenemase detection techniques [ 14 – 16 ], their implementation has been limited due to intricate procedures or high costs.

Recently, researchers have explored the potential of volatile organic compounds (VOCs) for strain identification [ 17 – 19 ], and applying volatile metabolites to antibiotic drug susceptibility testing has gained increasing attention [ 20 - 23 ]. For instance, gas chromatography-mass spectrometry (GC-MS) detection of 3-methyl-1-butanol enabled early identification of carbapenemase-positive CRKP strains [ 22 ], while non-targeted GC-MS analysis of VOC changes shows promise in the early CRKP strains identification [ 20 ]. Moreover, Smart et al. [ 23 ] employed thermal desorption-GC-MS to identify nine compounds that distinguished between cephalexin-sensitive and cephalexin-resistant isolates.

Gas chromatography-ion mobility spectrometry (GC-IMS) has emerged as a cutting-edge detection method in recent years. This innovative technique leverages the principles of chromatographic separation and ionization reactions to detect VOCs. The process begins with the pre-separation through a chromatographic column and is followed by their introduction into an ionization reaction zone by a carrier gas (nitrogen or air). Under the influence of an ion source, the carrier gas molecules and sample molecules undergo a series of ionization and ion-molecule reactions, resulting in the charging of sample molecules and their transformation into molecular ions. These ions are then driven by an electric field into a drift region through periodically opened ion gates, where they are separated and detected based on their varying migration rates due to continuous collisions with counterflowing neutral drift gas molecules [ 24 ]. Moreover, GC-IMS integrates the superior separation capabilities of gas chromatography (GC) with the enhanced sensitivity of ion mobility spectrometry (IMS). This innovative method eliminates the necessity for solid-phase microextraction by allowing the analysis of headspace components from solid or liquid samples through direct headspace injection. With a limit of detection reaching the parts per billion by volume (ppbv) level, GC-IMS facilitates both qualitative and quantitative analysis of individual compounds or labelers [ 25 , 26 ].

3-hydroxy-2-butanone (acetoin) emerges as a pivotal physiological metabolite synthesized by a diverse array of microorganisms thriving in glucose-enriched environments or other fermentable carbon sources. The catabolic conversion of this metabolite is primarily mediated by the acetoin dehydrogenase enzyme system (AoDH ES). Detecting acetoin-forming capacity, frequently achieved through the Voges-Proskauer reaction, serves as a valuable tool for classifying microorganisms. In bacteria grown in glucose-containing media, the release of 3-hydroxy-2-butanone (acetoin) is primarily regulated by the coordinated action of two enzymes: α-acetolactate synthase and α-acetolactate decarboxylase. 3-hydroxy-2-butanone (acetoin) plays a crucial role in these microorganisms, with its physiological significance encompassing acidification avoidance, nicotinamide adenine dinucleotide (NAD)/NADH (reduced form of NAD) ratio regulation, and carbon storage facilitation, ultimately maintaining the metabolic homeostasis and adaptability of microbial communities [ 27 , 28 ]. Notably, previous studies have demonstrated that K. pneumoniae releases 3-hydroxy-2-butanone (acetoin) during blood cultures [ 21 , 29 ] and in trypticase soy broth (TSB) [ 30 ].

Given the preceding justification and the escalating detection rate of CRKP in blood cultures [ 31 , 32 ], this study builds upon previous findings [ 21 ]. The previous research demonstrated the potential of GC-IMS for non-targeted analysis in simulated blood cultures, effectively identifying CRKP strains [ 21 ]. Therefore, this study aimed to utilize GC-IMS technology to detect a specific VOC biomarker, 3-hydroxy-2-butanone (acetoin), in antimicrobial-resistant KPC-producing bacteria, facilitating early identification of carbapenemase-producing CRKP strains.

Material and Methods

Ethics statement.

In this study, K. pneumoniae isolation was conducted following hospital laboratory protocol without identifiable patient data linked to the samples. Consequently, the Medical Research Ethics Committee of the Second Affiliated Hospital of Nanchang University exempted the study from ethics approval.

Sources of the Strains and Carbapenemase Detection

The standard strains of K. pneumoniae , including ATCC BAA-700603 (carbapenemase-negative), ATCC BAA-1706 (carbapenemase-negative), ATCC BAA-1705 ( bla KPC -positive), ATCC BAA-2146 ( bla NDM -positive), and ATCC BAA-2524 ( bla OXA -48 -positive), were obtained from the American Type Culture Collection (ATCC, USA).

Additionally, 69 clinical strains were isolated from the Second Affiliated Hospital of Nanchang University and are currently maintained in the research team’s strain bank [ 21 , 22 ]. Among these, 25 strains were carbapenem-susceptible K. pneumoniae (CSKP) strains, while the remaining 44 strains (CRKP) exhibited resistance to carbapenem. All K. pneumoniae strains (standard and clinical strains) underwent re-testing for antimicrobial susceptibility, interpreted according to the CLSI 2022 [ 12 ] guidelines. The mCIM and modified ethylenediaminetetraacetic acid (EDTA)-carbapenem inactivation method (eCIM) [ 12 ] were performed again to validate the carbapenemase type, confirmed by polymerase chain reaction (PCR). Moreover, all clinical isolates were evaluated for the presence of carbapenemase-related genes utilizing macrogenomic next-generation sequencing (mNGS) technology, provided by Qiantang Life Science Technology Co. Ltd. (Suzhou, China). Subsequently, all K. pneumoniae strains were stored at −80°C for further analysis.

Bacterial Culture and Sample Preparation

Figure 1 shows the growth curve of K. pneumoniae (ATCC BAA-700603) in blood culture bottle medium (BacT/ALERT ® SA; Ref. 259789; Biomérieux, Nürtingen, Germany), with an initial concentration of 10 7 colony-forming units per milliliter (CFU/mL) bacteria at 37°C and agitation of 200 revolutions per minute (rpm) in a 6 mL volume. The growth curve indicates that K. pneumoniae grew fastest after 3 h of incubation and reached the end of the exponential growth phase at around 5 h.

The growth curve of K. pneumoniae (ATCC BAA-700603). The free online tool Chiplot ( https://www.chiplot.online ). K. pneumoniae – Klebsiella pneumoniae; ATCC – American Type Culture Collection.

All experimental strains were inoculated on Columbia blood agar plates and incubated overnight at 37°C. Bacterial suspensions were prepared and added to test tubes containing 6 mL of blood culture bottle medium (initially 10 7 CFU/mL), capped, and incubated at 37°C with agitation at 200 rpm. Blank culture media served as control. To evaluate the effect of IPM on 3-hydroxy-2-butanone (acetoin) release from K. pneumoniae , a final concentration of 0.25 mg/mL of IPM (Solarbio, China) was added after 3 h (T0) of incubation. The final concentration of IPM in carbapenemase-negative CRKP strains was 16 μg/mL. To assess the impact of carbapenemase inhibitors on 3-hydroxy-2-butanone (acetoin) emission, IPM and avibactam sodium (1 mg/L) [ 33 ] or DPA (100 μg/mL) [ 34 ] were added after 3 h (T0). The standard strains and clinical strains were processed identically. Experiments were repeated six times for standard strains and in triplicate for clinical strains.

Following the inoculation of standard strains, with or without IPM, into a blood culture bottle medium, 500 μL samples were collected for GC-IMS analysis at various incubation periods: 3h (T0), 4h (T1), 5h (T2), 6h (T3), and 7h (T4). GC-IMS analysis of the standard strain with carbapenemase inhibitors and clinical strains was performed at the T2 time point.

The Measurement of 3-hydroxy-2-butanone (Acetoin) by GC-IMS

The FlavourSpec ® GC-IMS equipment (G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China) was employed to detect 3-hydroxy-2-butanone (acetoin). The GC-IMS parameters were set as follows: incubation time: 3 min; incubation temperature: 60°C; rotating speed: 500 rpm; injector temperature: 85°C; headspace gas volume: 1 mL; drift tube temperature (T1): 45°C; chromatographic column temperature (T2): 80°C; chromatographic column: MXT-WAX column (high-polar column, 15 m×0.53 mm, 0.1 μm, RESTEK, Bellefonte, PA, USA); carrier and drift gases: nitrogen (N 2 ) of 99.99% purity (Jiangzhu Industrial Co., Ltd., Nanchang, China); ionization source: tritium source; average radiation energy: 5.68 keV; drift tube length: 98 mm; ionization mode: positive ionization; drift tube flow: 150 mL/min; chromatographic column flow: 2 mL/min (0–3 min), 10 mL/min (3–10 min); total analysis time: 10 min.

The fingerprint of VOCs emitted by K. pneumoniae (ATCC BAA-700603) after 5 h of incubation in the medium of blood culture bottles (BacT/ALERT® SA) is depicted in Figure 2 , and the details of VOCs have been described in previous studies [ 21 ]. The VOC fingerprints generated by all the understudied K. pneumoniae strains, following a 5-h incubation period in a blood culture flask substrate, aligned with the pattern displayed in Figure 2 (without IPM addition).

The fingerprint of VOCs emitted by K. pneumoniae (ATCC BAA-700603) after 5 h incubation in blood culture bottles (BacT/ALERT ® SA) medium. Red boxes indicate VOCs emitted by K. pneumoniae, while blue boxes represent VOCs absorbed by K. pneumoniae . VOCal (version 0.1.3), G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China. VOCs – volatile organic compounds; K. pneumoniae – Klebsiella pneumoniae ; ATCC – American Type Culture Collection; M – monomer; D – dimer.

Using C4-C9 ketones (2-butanone, 2-pentanone, 2-hexanone, 2-heptanone, 2-octanone, 2-nonanone), from Shandong HaiNeng Scientific Instrument Co., Ltd. (Shandong, China), as reference standards for data calibration, 3-hydroxy-2-butanone (acetoin) was identified based on the retention index (RI) and drift time (RIP relative) in the GC-IMS library (NIST library and IMS library) [ 21 , 35 ]. GC-IMS detection of 3-hydroxy-2-butanone (acetoin) is characterized as follows: Chemical Abstract Service Registry Number (CAS#): C513860; Formula: C 4 H 8 O 2 ; Molecular weight (MW): 88.1; RI: 1304.2; Retention time (Rt): 226.606 s; and Drift time (Dt): 1.33401 a.u. (arbitrary units).

Statistical Analysis

The VOC fingerprint emitted by K. pneumoniae and the pseudo-3D plots of 3-hydroxy-2-butanone (acetoin) were generated by VOCal software (version 0.1.3; G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China). The relative peak volume values (integrating peak intensities within specific regions after comparing calibrations) of 3-hydroxy-2-butanone (acetoin) were extracted from the VOCal software and exported to Microsoft Excel (Excel for MacOS, 2020) for further analysis. All statistical analyses were performed using R (version 4.2.3; The R Foundation, Vienna, Austria). As the data did not follow a normal distribution, the Mann-Whitney U test was used for comparisons between groups . A P -value less than 0.05 was considered statistically significant ( P <0.05). Data visualizations were created using the online tool Chiplot ( https://www.chiplot.online ) [ 36 ].

Effect of IPM on the Release of 3-hydroxy-2-butanone (Acetoin) from Standard Strains of K. pneumoniae

The content of 3-hydroxy-2-butanone (acetoin) released by all standard strains increases progressively over time (T0–T4) without IPM addition ( Figure 3A ). However, after a 3-h incubation period (T0), IPM addition to the culture medium resulted in a consistent level of 3-hydroxy-2-butanone (acetoin) release by K. pneumoniae ATCC BAA-1706 (carbapenemase-negative), with no further changes in content observed from T0 to T4. In contrast, carbapenemase-positive standard strains displayed a temporal variation in 3-hydroxy-2-butanone (acetoin) release, similar to the same trend without IPM addition ( Figure 3B ). The content of 3-hydroxy-2-butanone (acetoin) exhibited a pronounced escalation between T1 and T2, aligning with the logarithmic phase of the growth curve ( Figure 1 ). Therefore, subsequent analysis focused mainly on the T2 time point.

The temporal variation patterns of 3-hydroxy-2-butanone (acetoin) emission by K. pneumoniae (standard strains, T0-T4). ( A ) The temporal variation patterns of 3-hydroxy-2-butanone (acetoin) emission by K. pneumoniae strains when imipenem was not added. ( B ) The temporal patterns of 3-hydroxy-2-butanone (acetoin) emitted by K. pneumoniae strains when imipenem was added. The free online tool Chiplot ( https://www.chiplot.online ). ATCC – American Type Culture Collection.

With IPM addition, the pseudo-3D plots illustrating the change of 3-hydroxy-2-butanone (acetoin) in standard strains are shown in Figure 4A (T2). Notably, IPM addition resulted in a significant decrease in 3-hydroxy-2-butanone (acetoin) contents emitted by K. pneumoniae ATCC BAA-1706 compared to its release without IPM. Interestingly, no statistically significant difference was observed in the relative peak volume (semi-quantified) of 3-hydroxy-2-butanone (acetoin) emitted by K. pneumoniae ATCC BAA-1706 at T2 with IPM addition, compared to the blank medium ( P >0.05) ( Figure 4B ).

The changes in 3-hydroxy-2-butanone (acetoin) contents emitted by standard strain at the T2 time point before and after the addition of imipenem. ( A ) The pseudo-3D plots of 3-hydroxy-2-butanone (acetoin) in each standard strain (comparison between no addition of imipenem and imipenem addition). VOCal (version 0.1.3), G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China. ( B ) Comparison of the relative peak volume (semiquantitative) of 3-hydroxy-2-butanone (acetoin) emitted by standard strains (with imipenem and without imipenem). The free online tool Chiplot ( https://www.chiplot.online ). The Mann-Whitney U test was used for pairwise comparison. * P <0.05, ** P <0.01, *** P <0.001, and **** P <0.0001. R version 4.2.3, The R Foundation, Vienna, Austria. ATCC – American Type Culture Collection; IPM – imipenem.

Effect of Carbapenemase Inhibitors on the Release of 3-hydroxy-2-butanone (Acetoin) from Standard Strains of K. pneumoniae

The effect of carbapenemase inhibitors on the content of 3-hydroxy-2-butanone (acetoin) emitted by K. pneumoniae was further investigated. Figure 5A shows the pseudo-3D plots illustrating the changes in 3-hydroxy-2-butanone (acetoin) content following the introduction of carbapenemase inhibitors in various standard strains. The combination of IPM and avibactam sodium significantly decreased 3-hydroxy-2-butanone (acetoin) content in K. pneumoniae ATCC BAA-1705 ( bla KPC -positive), compared to IPM alone or IPM with DPA (both P <0.05; Figure 5B ). Similarly, the combination of IPM and DPA significantly decreased the 3-hydroxy-2-butanone (acetoin) content in K. pneumoniae ATCC BAA-2146 ( bla NDM -positive), compared to IPM alone or IPM with avibactam sodium (both P <0.05; Figure 5B ). However, compared with the addition of IPM alone, the remaining standard strains showed no notable changes in 3-hydroxy-2-butanone (acetoin) contents after introducing carbapenemase inhibitors (despite P <0.05, but no obvious visual changes observed in the pseudo-3D plots).

The changes in the content of 3-hydroxy-2-butanone (acetoin) emitted by standard strain at the T2 time point before and after the addition of carbapenemase inhibitors. ( A ) The pseudo-3D plots of 3-hydroxy-2-butanone (acetoin) contents in standard strains before and after the addition of carbapenase inhibitors. VOCal (version 0.1.3), G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China. ( B ) Comparison of the relative peak volume (semiquantitative) of 3-hydroxy-2-butanone (acetoin) emitted by standard strains before and after the addition of carbapenase inhibitors. The free online tool Chiplot ( https://www.chiplot.online ). The Mann-Whitney U test was used for pairwise comparison. * P <0.05. R version 4.2.3, The R Foundation, Vienna, Austria. ATCC – American Type Culture Collection; IPM – imipenem; DPA – pyridine-2,6-dicarboxylic acid.

The Potential Application Value of 3-hydroxy-2-butanone (Acetoin) in the Identification of Carbapenemase-Producing K. pneumoniae in Clinical Strains

A validation study utilizing clinical strains was conducted to further elucidate the role of 3-hydroxy-2-butanone (acetoin) in identifying carbapenemase-producing K. pneumoniae . As described in Table 1 , among 44 CRKP isolates, the mNGS examination discovered five carbapenemase-negative CRKP strains that were devoid of carbapenemase genes. Subsequently, based on the results of mCIM, eCIM, PCR and mNGS, these CRKP isolates were further classified based on carbapenemase type: KPC-positive CRKP strains (n=20), NDM-positive CRKP strains (n=15), imipenem-hydrolyzing β-lactamase (IMP)-positive CRKP strains (n=4), and carbapenemase-negative CRKP strains (n=5).

Effects of different treatments on 3-hydroxy-2-butanone content (relative peak volume, mean) at the T2 time point for each CRKP clinical isolate.

“−” represents the absence of carbapenemase or carbapenemase gene. IPM – imipenem; DPA – pyridine-2,6-dicarboxylic acid; CRKP – carbapenem-resistant Klebsiella pneumoniae; KPC – Klebsiella pneumoniae -carbapenemases; NDM – New Delhi metallo-β-lactamase; IMP – imipenemase metallo-β-lactamase; MIC – minimum inhibitory concentration.

The addition of IPM resulted in higher 3-hydroxy-2-butanone (acetoin) in the KPC-positive, NDM-positive, and IMP-positive CRKP groups compared to the CSKP group (n=25) (all P <0.05; Figure 6A, 6B ). However, the carbapenemase-negative CRKP group did not show a statistically significant difference in 3-hydroxy-2-butanone (acetoin) content when compared to the CSKP group ( P >0.05; Figure 6 ).

After imipenem addition, changes in the 3-hydroxy-2-butanone (acetoin) content emitted by various carbapenase types of K. pneumoniae strains (clinical isolates) at the T2 time point. ( A ) The pseudo-3D plots of 3-hydroxy-2-butanone (acetoin) in different carbapenase types of K. pneumoniae strains after the addition of imipenem. VOCal (version 0.1.3), G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China. ( B ) Comparison of the relative peak volume (semiquantitative) of 3-hydroxy-2-butanone (acetoin) emitted by different carbapenase types of K. pneumoniae strains after the imipenem addition. The free online tool Chiplot ( https://www.chiplot.online ). The Mann-Whitney U test was used for pairwise comparison. **** P <0.0001. R version 4.2.3, The R Foundation, Vienna, Austria. K. pneumoniae – Klebsiella pneumoniae ; CSKP – carbapenem-susceptible Klebsiella pneumoniae ; CRKP – carbapenem-resistant Klebsiella pneumoniae ; KPC – Klebsiella pneumoniae -carbapenemases; NDM – New Delhi metallo-β-lactamase; IMP – imipenemase metallo-β-lactamase.

The addition of carbapenemase inhibitors resulted in pseudo-3D plots similar to those of K. pneumoniae ATCC BAA-1705 for the KPC-positive group and K. pneumoniae ATCC BAA-2146 for the NDM-positive and IPM-positive groups ( Figure 7A ). Statistical analysis revealed a significant decrease in the relative peak volume of 3-hydroxy-2-butanone (acetoin) in the KPC-positive group with avibactam sodium addition and in the NDM-positive and IMP-positive groups with the DPA addition.

The changes in 3-hydroxy-2-butanone (acetoin) content emitted by different carbapenase types of K. pneumoniae strains (clinical isolates) at the T2 time point before and after the addition of carbapenase inhibitors. ( A ) The pseudo-3D plots of 3-hydroxy-2-butanone (acetoin) in different carbapenase types of K. pneumoniae strains before and after the addition of carbapenase inhibitors. VOCal (version 0.1.3), G.A.S., Shandong HaiNeng Scientific Instrument Co., Ltd., Shandong, China. ( B ) Comparison of the relative peak volume (semiquantitative) of 3-hydroxy-2-butanone (acetoin) emitted by different carbapenase types of K. pneumoniae strains before and after the addition of carbapenase inhibitors. The free online tool Chiplot ( https://www.chiplot.online ). The Mann-Whitney U test was used for pairwise comparison. * P <0.05, **** P <0.0001. R version 4.2.3, The R Foundation, Vienna, Austria. K. pneumoniae – Klebsiella pneumoniae ; CSKP – carbapenem-susceptible Klebsiella pneumoniae; CRKP – carbapenem-resistant Klebsiella pneumoniae; KPC – Klebsiella pneumoniae-carbapenemases; NDM – New Delhi metallo-β-lactamase; IMP – imipenemase metallo-β-lactamase; IPM – imipenem; DPA – pyridine-2,6-dicarboxylic acid.

Finally, Table 1 summarizes the carbapenemase types and genes of the clinical CRKP isolates, the minimum inhibitory concentration (MIC) of IPM, and the GC-IMS detection results of 3-hydroxy-2-butanone (acetoin) released by each CRKP strain under three different treatments. It is noteworthy that the trends of 3-hydroxy-2-butanone (acetoin) content under different treatments for each CRKP strain were consistent with the results above.

This study expanded previous research [ 22 ] by utilizing a more sensitive GC-IMS assay and blood culture bottles (BacT/ALERT ® SA) commonly used in clinical settings as the medium. Additionally, the study investigated temporal variation patterns in the distinctive VOC, 3-hydroxy-2-butanone (acetoin). The findings suggest that 3-hydroxy-2-butanone (acetoin) has the potential as a biomarker for identifying carbapenemase-producing K. pneumoniae , supported by the following pieces of evidence: (1) The temporal release pattern of 3-hydroxy-2-butanone aligns with the observed bacterial growth pattern of K. pneumoniae in blood culture matrix. (2) IPM addition resulted in a significant decrease in 3-hydroxy-2-butanone (acetoin) release by carbapenemase-negative strains during the late exponential growth phase (T2), in comparison to carbapenemase-positive strains (both standard and clinical strains). (3) The addition of carbapenemase inhibitors enabled further identification of carbapenemase phenotypes (class A or B) based on changes in 3-hydroxy-2-butanone (acetoin) content.

Research has shown that K. pneumoniae can release 3-hydroxy-2-butanone (acetoin) during its growth process [ 29 , 30 ]. As a fermentation bacterium, K. pneumoniae phosphorylates extracellular glucose and other monosaccharides into glucose 6-phosphate through the phosphotransferase system (PTS) [ 37 , 38 ]. This glucose 6-phosphate is then utilized by the bacteria through the glycolytic pathway, producing pyruvate, which is further metabolized to generate acetolactate. The enzyme α-acetolactate decarboxylase ultimately converts acetolactate to 3-hydroxy-2-butanone (acetoin) [ 27 , 30 , 39 ]. Although the metabolic mechanism of 3-hydroxy-2-butanone (acetoin) in K. pneumoniae requires further investigation, the findings of this study demonstrated that IPM significantly reduced 3-hydroxy-2-butanone (acetoin) release from CSKP as compared to the absence of IPM. IPM probably exerts its inhibitory effect on bacterial cell wall synthesis, resulting in bacterial cell wall defects, subsequent bacterium expansion, and cell lysis due to the alteration of bacterial cytoplasmic osmotic pressure, ultimately causing bacterial death [ 40 ]. Apparently, CSKP strains were killed by IPM, and their growth metabolism ceased, resulting in significantly reduced release of 3-hydroxy-2-butanone (acetoin), which is closely related to the growth metabolism of K. pneumoniae , consistent with previous research [ 20 ].

Moreover, avibactam, a newly developed β-lactamase inhibitor, possesses a non-lactam structural scaffold and lacks substantial antimicrobial efficacy independently. However, it can effectively inhibit class A (including ESBLs and KPCs), class C, and some class D β-lactamases [ 41 ]. Notably, adding avibactam sodium significantly reduced 3-hydroxy-2-butanone (acetoin) release in class A carbapenemase-producing K. pneumoniae but had no significant effect on other strains compared to no addition of avibactam sodium. Previous studies have demonstrated the inhibitory impact of avibactam sodium on OXA-48 carbapenemase [ 42 , 43 ], but our findings did not replicate this, possibly attributed to a low concentration of avibactam sodium. Further investigation is needed to ascertain the underlying cause. Meanwhile, DPA, a novel metallo-β-lactamase effective against class B carbapenemases [ 44 ], selectively hindered 3-hydroxy-2-butanone (acetoin) release in K. pneumoniae strains producing class B carbapenemases (NDM- and IMP-positive strains), as observed in this study.

The findings from the previous study [ 22 ] indicated that 3-methyl-1-butanol has the potential to serve as a biomarker for the detection of carbapenemase-producing K. pneumoniae in TSB. However, further analysis using pseudo-3D plots and semiquantitative analysis in this study revealed that alteration in 3-hydroxy-2-butanone (acetoin) exhibited greater specificity. Consequently, 3-methyl-1-butanol was not subjected to further investigation in this study. In fact, this discrepancy can be attributed to variations in the culture substrates and assay methods employed. Meanwhile, one of the previous studies [ 21 ] successfully identified CRKP strains by analyzing changes in K. pneumoniae -associated VOCs under different treatments, but this approach was non-targeted and could not identify characteristic VOCs. Therefore, the present study improves upon this approach and simplifies the detection of carbapenemase-producing K. pneumoniae by utilizing the characteristic alterations of 3-hydroxy-2-butanone (acetoin) as an entry point.

Unfortunately, the limited diversity of clinical strains in our collection, predominantly KPC-positive CRKP strains, restricted the complete validation of current findings, necessitating a cautious interpretation of results. However, epidemiological data revealed that bla KPC-2 emerged as the prevailing carbapenemase gene in China, accounting for up to 94% of the clinical CRKP isolates [ 31 ], consistent with the regional epidemiological data [ 22 ]. Notably, the utilization of pseudo-3D plots and semiquantitative analysis of the content of 3-hydroxy-2-butanone (acetoin) in this study facilitated the accurate identification of CRKP strains that produce class A or B carbapenemases, providing a valuable reference for future investigations.

Another point of concern was the identification of carbapenemase-negative CRKP. These strains cannot produce carbapenemases and owe their carbapenem resistance to the absence of membrane pore proteins OmpK35 and OmpK36, combined with the expression of extended-spectrum β-lactamase enzymes ( bla CTX-M , bla SHV , and bla AmpC ) or high expression of efflux pumps [ 45 ]. Moreover, the non-uniform minimum inhibitory concentrations (MIC) of these strains, despite their IPM resistance, hindered the distinction between CSKP and carbapenemase-negative CRKP based on 3-hydroxy-2-butanone (acetoin) changes, even at a reduced imipenem concentration of 16 μg/mL. Therefore, future studies require a more comprehensive investigation into the optimal imipenem concentration to address this challenge.

The present study has several strengths: (1) Commercial blood culture bottles were used as the medium to enhance clinical applicability, providing a new strategy for early identification of carbapenemase-producing CRKP bacteria, unlike traditional media such as TSB [ 20 , 22 ]. (2) GC-IMS offers superior trace detection capabilities, achieving detection limits as low as nanograms or even picograms [ 46 ], and offers advantages over GC-MS, including reduced detection time, no pre-treatment requirements, and rapid sample detection (within 13 min in our study), thus facilitating batch detection. GC-IMS also employs fingerprinting and pseudo-3D plots for intuitive visualization of VOC disparities. (3) The study demonstrates the potential of 3-hydroxy-2-butanone (acetoin) for identifying carbapenemase-producing CRKP in the stromal fluid of blood culture bottles, using standard strains as references and clinical strains for validation, highlighting its significance in this context.

Limitations

This study has certain limitations. First, the impact of blood addition on 3-hydroxy-2-butanone (acetoin) release remains uncertain, consistent with the previous studies [ 30 ] indicating that blood addition has no influence on the VOC content produced by K. pneumoniae in blood-free media. This phenomenon warrants further research to understand its impact on VOC contents. Second, the exclusive use of clinical strains from a single medical center limits the generalizability of our method, highlighting the need for larger sample sizes and diverse carbapenemases to reinforce our research outcomes. Third, the GC-IMS assay technique, while valuable, requires specialized instrumentation and expertise and incurs higher costs, making it less accessible in the medical field. Fourth, detecting carbapenemase-positive strains based on the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) method is recommended by the CLSI guidelines [ 12 ]. Regrettably, this study did not employ this method because: 1) Although MALDI-TOF MS technology can rapidly detect the types of carbapenemases, its disadvantage lies in the fact that there are currently only a limited number of studies focusing on specific carbapenemases, and there is no unified and feasible methodology [ 47 ]. 2) Although MALDI-TOF MS technology has established a relatively comprehensive database for microbial identification, the existing databases may still be incomplete or have accuracy issues when detecting specific strains or resistance patterns. This may lead to misdiagnosis or missed detection of carbapenemase-producing strains. Instead, this research proposed GC-IMS as a new strategy for detecting carbapenemase-positive CRKP strains by identifying changes in 3-hydroxy-2-butanone (acetoin). This method complements the MALDI-TOF MS-based method and aims to improve the speed and accuracy of detecting carbapenemase-positive CRKP strains. We believe that combining both methods can effectively combat the spread of drug-resistant strains, ensuring patient health and safety. Lastly, further evaluation is needed to confirm the specificity of 3-hydroxy-2-butanone (acetoin) for K. pneumoniae ; however, utilizing a fingerprint, as depicted in Figure 2 , will aid in differentiating K. pneumoniae from other bacterial species.

In conclusion, GC-IMS was successfully utilized to identify carbapenemase-producing K. pneumoniae through changes in 3-hydroxy-2-butanone (acetoin) content after IPM addition. Simultaneously, the inclusion of carbapenemase inhibitors enabled the differentiation between class A and B carbapenemases. However, further research is warranted to confirm these findings.

Acknowledgments

The authors would like to thank Home for Researchers ( www.home-for-researchers.com ) for English language editing services. We also acknowledge engineer Liqiang Zhao for helping with GC-IMS data analysis.

Conflict of interest: None declared

Declaration of Figures’ Authenticity: All figures submitted are original creations and have not been previously published or duplicated, either in whole or in part. They appear in this work for the first time.

Financial support: This work was supported by the National Natural Science Foundation of China (82060391), the Postgraduate Innovation Special Foundation of Jiangxi Province (YC2023-B093), the Natural Science Foundation of Jiangxi Province (20202BAB216021), the Medical Health Science and Technology Project of Jiangxi Provincial Health Commission (20201034 and 202130412), and the Chinese Medical Science and Technology Research Projects of Jiangxi Provincial Administration of Traditional Chinese Medicine (2023Z030).

• Analytical Chemistry

Column Chromatography

What is column chromatography.

In chemistry, Column chromatography is a technique which is used to separate a single chemical compound from a mixture dissolved in a fluid.

Column chromatography separates substances based on differential adsorption of compounds to the adsorbent as the compounds move through the column at different rates which allows them to get separated in fractions. This technique can be used on a small scale as well as large scale to purify materials that can be used in future experiments. This method is a type of adsorption chromatography technique.

Table of Content

• Applications
• Frequently Asked Questions – FAQs

Column Chromatography Principle

When the mobile phase along with the mixture that needs to be separated is introduced from the top of the column, the movement of the individual components of the mixture is at different rates. The components with lower adsorption and affinity to the stationary phase travel faster when compared to the greater adsorption and affinity with the stationary phase. The components that move fast are removed first whereas the components that move slowly are eluted out last.

The adsorption of solute molecules to the column occurs in a reversible manner. The rate of the movement of the components is expressed as:

R f = the distance travelled by solute/ the distance travelled by the solvent

R f is the retardation factor.

Column Chromatography Diagram

Elution is a chemical process that involves removing a material’s ions by ion exchange with another material. The chromatographic technique of extracting an adsorbed substance from a solid adsorbing media using a solvent. The eluent is the solvent or mobile phase that passes through the column. When the polarity of the eluent matches the polarity of the molecules in the sample, the molecules desorb from the adsorbent and dissolve in the eluent.

The fraction of the mobile phase that transports the sample components is known as eluent. The mixture of solute and solvent that exits the column is known as an eluate. The eluate is made up of the mobile phase and analytes. A substance that separates and moves constituents of a mixture through the column of a chromatograph. The eluent in liquid chromatography is a liquid solvent whereas in gas chromatography is a carrier gas.

Column Chromatography Procedure

Before starting with the Column Chromatography Experiment let us understand the different phases involved.

Mobile phase – This phase is made up of solvents and it performs the following functions:

• It acts as a solvent-sample mixture that can be introduced in the column.
• It acts as a developing agent – helps in the separation of components in the sample to form bands.
• It acts as an eluting agent – the components that are separated during the experiment are removed from the column
• Some examples of solvents used as mobile phases based on their polarity are – ethanol, acetone, water, acetic acid , pyridine, etc.

Stationary phase – It is a solid material which should have good adsorption properties and meet the conditions given below:

• Shape and size of particle: Particles should have a uniform shape and size in the range of 60 – 200μ in diameter.
• Stability and inertness of particles: high mechanical stability and chemically inert. Also, no reaction with acids or bases or any other solvents was used during the experiment.
• It should be colourless, inexpensive and readily available.
• Should allow free flow of mobile phase
• It should be suitable for the separation of mixtures of various compounds.

Column Chromatography Experiment

• The stationary phase is made wet with the help of solvent as the upper level of the mobile phase and the stationary phase should match. The mobile phase or eluent is either solvent or a mixture of solvents. In the first step the compound mixture that needs to be separated, is added from the top of the column without disturbing the top level. The tap is turned on and the adsorption process on the surface of silica begins.
• Without disturbing the stationary phase solvent mixture is added slowly by touching the sides of the glass column. The solvent is added throughout the experiment as per the requirement.
• The tap is turned on to initiate the movement of compounds in the mixture. The movement is based on the polarity of molecules in the sample. The non-polar components move at a greater speed when compared to the polar components.
• For example, a compound mixture consists of three different compounds viz red, blue, green then their order based on polarity will be as follows blue>red>green
• As the polarity of the green compound is less, it will move first. When it arrives at the end of the column it is collected in a clean test tube. After this, the red compound is collected and at last blue compound is collected. All these are collected in separate test tubes.

Column Chromatography Applications

• Column Chromatography is used to isolate active ingredients.
• It is very helpful in separating compound mixtures.
• It is used to determine drug estimation from drug formulations.
• It is used to remove impurities.
• Used to isolate metabolites from biological fluids.

Types of Column Chromatography:

1. Adsorption column chromatography – Adsorption chromatography is a technique of separation, in which the components of the mixture are adsorbed on the surface of the adsorbent.

2. Partition column chromatography – The stationary phase, as well as mobile phase, are liquid in partition chromatography .

3. Gel column chromatography – In this method of chromatography, the separation takes place through a column packed with gel. The stationary phase is a solvent held in the gap of a solvent.

Frequently Asked Questions – FAQs

What is the principle involved in column chromatography.

The basic principle involved in column chromatography is to adsorb solutes of the solution with the help of a stationary phase and further separate the mixture into discrete components.

What is column chromatography?

It is a precursory technique used in the purification of compounds based on their hydrophobicity or polarity. In this chromatography process, the molecule mixture is separated depending on its differentials partitioning between a stationary phase and a mobile phase.

What is the main advantage of column chromatography?

The main advantage of this chromatography technique is that the stationary phase is less expensive and can be easily disposed of as it undergoes recycling.

How are the compounds separated in this technique?

The separation is similar to that of TLC where the compound mixture is carried by a mobile phase via a stationary phase.

Which compounds elute out first in the column chromatography technique?

Non-polar compounds. The polar compounds will strongly commune with the silica when compared to the non-polar compounds.

What is elution in column chromatography?

What are the limitations of column chromatography, what are the different types of column chromatography, what is an adsorption column chromatography, what is gel column chromatography.

To learn more about the different types of Column Chromatography from the experts, register to BYJU’S now!

Other important links:

Put your understanding of this concept to test by answering a few MCQs. Click ‘Start Quiz’ to begin!

Select the correct answer and click on the “Finish” button Check your score and answers at the end of the quiz

Visit BYJU’S for all Chemistry related queries and study materials

Your result is as below

Request OTP on Voice Call

Leave a Comment Cancel reply

Your Mobile number and Email id will not be published. Required fields are marked *

Post My Comment

What is developing agent?

Chromatographic procedure generally involves introducing at the top of the column the mixture of the components to be separated, developing the mixture with a suitable agent, and collecting the components in separate effluent fractions. The best developing agent in thin-layer chromatography was Petroleum ether: ethyl acetate.

Register with BYJU'S & Download Free PDFs

Register with byju's & watch live videos.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 10 August 2024

Covalent penicillin-protein conjugates elicit anti-drug antibodies that are clonally and functionally restricted

• Lachlan P. Deimel   ORCID: orcid.org/0000-0003-3803-871X 1   nAff9 ,
• Lucile Moynié 2 ,
• Guoxuan Sun 2 ,
• Viliyana Lewis   ORCID: orcid.org/0000-0002-4882-336X 2 ,
• Abigail Turner   ORCID: orcid.org/0000-0002-3995-6558 2 ,
• Charles J. Buchanan   ORCID: orcid.org/0000-0002-1840-5706 2 , 3 , 4 ,
• Sean A. Burnap 4 , 5 ,
• Mikhail Kutuzov   ORCID: orcid.org/0000-0003-3386-4350 1 ,
• Carolin M. Kobras 1 ,
• Yana Demyaneko   ORCID: orcid.org/0000-0002-6628-1912 2 , 6 ,
• Shabaz Mohammed   ORCID: orcid.org/0000-0003-2640-9560 2 , 3 , 5 ,
• Mathew Stracy 1 ,
• Weston B. Struwe   ORCID: orcid.org/0000-0003-0594-226X 4 , 5 ,
• Andrew J. Baldwin   ORCID: orcid.org/0000-0001-7579-8844 2 , 3 , 4 ,
• James Naismith   ORCID: orcid.org/0000-0001-6744-5061 2 ,
• Benjamin G. Davis   ORCID: orcid.org/0000-0002-5056-407X 2 , 3 , 6 &
• Quentin J. Sattentau   ORCID: orcid.org/0000-0001-7170-1937 1 , 7 , 8  

Nature Communications volume  15 , Article number:  6851 ( 2024 ) Cite this article

5 Altmetric

Metrics details

• Chemical biology
• NMR spectroscopy
• Structural biology

Many archetypal and emerging classes of small-molecule therapeutics form covalent protein adducts. In vivo, both the resulting conjugates and their off-target side-conjugates have the potential to elicit antibodies, with implications for allergy and drug sequestration. Although β-lactam antibiotics are a drug class long associated with these immunological phenomena, the molecular underpinnings of off-target drug-protein conjugation and consequent drug-specific immune responses remain incomplete. Here, using the classical β-lactam penicillin G (PenG), we probe the B and T cell determinants of drug-specific IgG responses to such conjugates in mice. Deep B cell clonotyping reveals a dominant murine clonal antibody class encompassing phylogenetically-related IGHV1 , IGHV5 and IGHV10 subgroup gene segments. Protein NMR and x-ray structural analyses reveal that these drive structurally convergent binding modes in adduct-specific antibody clones. Their common primary recognition mechanisms of the penicillin side-chain moiety (phenylacetamide in PenG)—regardless of CDRH3 length—limits cross-reactivity against other β-lactam antibiotics. This immunogenetics-guided discovery of the limited binding solutions available to antibodies against side products of an archetypal covalent inhibitor now suggests future potential strategies for the ‘germline-guided reverse engineering’ of such drugs away from unwanted immune responses.

Similar content being viewed by others

The hinge-engineered IgG1-IgG3 hybrid subclass IgGh47 potently enhances Fc-mediated function of anti-streptococcal and SARS-CoV-2 antibodies

Crystal structure of ultra-humanized anti-pTau Fab reveals how germline substitutions humanize CDRs without loss of binding’

A deimmunized and pharmacologically optimized Toll-like receptor 5 agonist for therapeutic applications

Introduction.

In isolation, non-protein, low molecular weight compounds are typically non-immunogenic to the mammalian immune system. As exemplified by classical hapten-carrier biology, antibody responses against small molecules such as 4-hydroxy-3-nitrophenol acetyl (NP) require conjugation to a suitable carrier protein 1 . However, upon attachment to protein, resulting epitopes arise within an antigenic complex that may cross-link cognate B cell receptors (BCRs) and that are associated with peptidic components that may be presented to T helper (Th) cells; together, these can impart combined help to propagate a specific B cell population to a given attached compound 2 , 3 , 4 .

In principle, these mechanisms may extend to small-molecule drugs, particularly those with reactive functional groups that drive covalent conjugation with endogenous proteins under physiological conditions 5 , 6 . Unwanted immune responses to covalent bond-forming drugs, particularly in the form of anti-drug antibody (ADA) responses, include hypersensitivity and allergy reactions, the most severe of which can be life-threatening. Whilst the number of covalent bond-forming drugs (e.g. covalent inhibitors) in clinical use has been traditionally limited 7 , in recent years there has been strong renewed interest 8 , 9 yet notably little analysis of unwanted drug-immune system interactions.

The best-characterised examples of unwanted immunogenicity from drug-protein conjugates are β-lactam antibiotics, such as penicillin G (PenG) 10 . As the electrophilic source of its inhibitory activity, the β-lactam group of PenG may also drive background / side reactivity with off-target biological nucleophiles leading to protein conjugation via primary amine-containing sidechains of lysine (and potentially other nucleophilic residues including arginine, histidine, and cysteine), as has been observed under some buffer conditions 11 , 12 . Such protein-PenG complexes are the antigenic determinants of antibiotic hypersensitivity. The mechanistic underpinnings of the hypersensitivity reaction are immunologically heterologous, with the most common and well characterised being T helper (Th) cell-mediated (type IV) that may be elicited in up to 30% of the population 13 , 14 , 15 , 16 , 17 . However, the most clinically severe forms of drug hypersensitivity are antibody-mediated, particularly IgE-induced anaphylaxis 18 . IgG-mediated hypersensitivity is less severe but relatively common 18 , 19 .

Penicillin is one of the most frequent causes of anaphylaxis and anaphylaxis-related deaths in humans 19 . However, penicillin allergy diagnosis is currently highly inaccurate. Nearly 6% of the general population in the UK are recorded as having a penicillin allergy, yet more than 95% of these patients can ultimately tolerate this class of drug, indicating that most patients are falsely recorded as allergic 20 . Patients with a penicillin allergy record have an increased risk of Clostridioides difficile and Methicillin-resistant Staphylococcus aureus infections and death; this is presumably through increased use of alternatives to β-lactam antibiotics 21 . Furthermore, penicillin allergy diagnosis is associated with higher numbers of total antibiotic prescriptions 22 , undermining antimicrobial stewardship goals and increasing the risk for antimicrobial resistance 23 . A better understanding of the immunological basis of penicillin hypersensitivity is therefore vitally needed to help predict which antibiotic recipients are, or will become, allergic 24 , 25 , and to inform potential future deleterious immune reactions against new generations of covalent bond-forming drugs.

Notably, although the first descriptions of penicilloyl-directed serological responses were made in 1961 12 , key phenomena remain incompletely understood, including (i) the biochemical basis of PenG–protein adduction in vivo and in vitro; (ii) the relative immunogenicity of fully chemically characterised and purified PenG adducts; (iii) the immunophylogenetics of B cells specific to PenG-protein complexes; and (iv) the structure/function characteristics of antibody clones specific to these adducts.

Here, through systematic complementary biochemical, structural and clonotypic analyses of the relationship between the protein-conjugating properties of PenG and its immunogenicity in a mouse model, we now fully map the PenG-specific antibody response. We find that the ADA response is based upon a restricted cluster of highly related B cell germline clonal families that, regardless of CDRH3 length, engage the penicillin sidechain via conserved binding modes. These findings offer a rational basis for understanding ADA responses, and suggest that antibodies have limited binding solutions that in turn may inform drug ‘reverse engineering’ to avoid ADA.

Immunogenicity of ‘pre-complexed’ penicillin-protein conjugate antigen

PenG has constituent β-lactam, thiazolidine and benzylamide sidechain moieties (Fig.  1a ); the β-lactam ring is long known to react with nucleophilic protein sidechains, including the off-target ε-amino groups of lysine residues leading to the formation of subsequent ε-amide-drug adduct forms such as those formed via β-lactam ring opening 3 , 11 , 12 , 26 , 27 , 28 (Fig.  1b ). To probe PenG-protein off-target conjugation, we first titrated adduct formation on the model protein hen egg lysozyme (HEL, ~14 kDa, 129 a.a.), chosen for its stability and relatively evenly distributed (six) lysine (Lys) residues. Various buffer, drug amounts and pH conditions were tested in vitro, and global site-specific drug occupancy was evaluated via mass spectrometry (MS). These revealed pH- and buffer-modulated conjugation levels (Fig  S1 a; Document  S1 ). Mapping of adduct formation through tryptic digest followed by site-specific liquid chromatography tandem MS (LC-MS/MS) analysis confirmed adduct formation at all lysine residues 1, 13, 33, 96/97 and 116 (Fig  S1b ). These reaction data informed ex vivo conjugation of PenG to various recombinant carrier proteins at close-to-physiological pH for subsequent immunisation.

a Chemical structure of PenG. b Proposed target residues of electrophilic β-lactam; primary reported target of lysine with potential targeting of arginine and histidine. c 1 mg/mL HSA and PenG (1:200 per Lys) were mixed in vitro with 0.1 M HCO 3 -, pH = 8.0. This was left at 25 °C for 16 h before dialysis into PBS. d Sex-matched 6-week-old naïve WT C57BL/6 mice were twice immunized (wk 0 and 4) with 10 µg HSA or HSA-PenG in alum. Created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license. e Terminal HEL-specific IgG EPT was evaluated. PenG-specific endpoint titres were evaluated by screening cross-reactivity against HEL-PenG. IgG titres against PenG were evaluated both ( f ) longitudinally and ( g ) at the terminal timepoint. h Competition ELISA was conducted, wherein HSA-PenG antisera binding for HEL-PenG was competed out with soluble PenG. i Longitudinal protein backbone-specific, HSA, and ( j ) terminal HSA-PenG-specific IgG endpoint titres were screened. e – j Dots represent data from a single animal ( n  = 4 per group), and bars/text denotes the median (ND = not detected). Groups were compared via Mann–Whitney test (two-sided). Source data are provided as a Source Data file.

Pilot immunogenicity analysis of HEL-PenG conjugates formulated in aluminium hydroxide (alum) adjuvant was conducted by subcutaneous (s.c.) administration to C57BL/6 mice. Antisera were titrated by ELISA on an unrelated PenG-modified protein (human serum albumin-PenG; HSA-PenG) to determine the titres specifically against the penicilloyl adduct. This revealed modest but significant ( P  < 0.05 Mann–Whitney U) isotype-switched IgG responses raised against the drug adduct (Fig  S1d–g ).

Although a useful model antigen for biochemical characterisation and pilot immunogenicity analysis, HEL is a weak Th cell antigen in mice 29 , 30 . We therefore next evaluated the antibody response generated by PenG pre-complexed to more antigenic HSA using the conditions optimised for HEL (Fig.  1c ). Site-specific occupancy of penicilloyl adducts was evaluated via LC-MS/MS and again diverse lysine occupancy was observed (Fig  S2 ). These occupancy data are concordant with previously published drug modification sites of HSA 28 , 31 , 32 , 33 . Mice were immunised with HSA or HSA-PenG formulated in alum, followed by periodic blood sampling (Fig.  1d ). Anti-penicilloyl serum IgG responses were measured by ELISA against HEL-PenG; no IgG cross-reactivity was detected against unmodified HEL (Fig.  1e ) . However, strikingly, post-prime HSA-PenG antisera displayed considerable IgG reactivity with HEL-PenG, whereas no reactivity was detected in the HSA-alone antiserum (Fig.  1f ). Together, these immediately suggested an anti-PenG-adduct-specific response. Post-boost and at the terminal timepoint, the median HEL-PenG-specific IgG endpoint titre (EPT) was marked at ~2.2 × 10 5 for the HSA-PenG antisera and a near-baseline EPT of ~1.7 × 10 2 for the control HSA antisera ( P  = 0.029, Mann-Whitney test) (Fig.  1g ). Analysis of CD4 + Th responses revealed significant T cell proliferation and IFN-γ production only in the HSA-PenG-restimulated cells, consistent with the adducted protein being most efficiently captured and processed by B cells and presented to Th cells (Fig  S3 ).

The observed antibody responses were generated against a protein-conjugated PenG derivative. Characterisation was consistent with direct β-lactam opening, but we cannot discount a pathway involving intermediates of penicillanic acid (Document  S1 ) 27 , 28 . HSA-PenG antiserum binding to HEL-PenG was out-competed by free PenG, with a median IC 50 of 1.8 mM, while kanamycin (an unrelated non-β-lactam-type antibiotic) failed to detectably compete for antibody binding (Fig.  1h ). These data reveal that antibodies raised against the autologous PenG adduct are cross-reactive with free PenG, suggesting recognition of a common motif that is not the ‘opened’ β-lactam. Administration of HSA-PenG did not affect the antibody responses against the protein unmodified protein backbone, compared with mice immunised with HSA alone ( P  > 0.9999, Mann–Whitney test) (Fig.  1i,j ).

Self-protein carriers elicit penicillin-specific antibodies

Having demonstrated strong B cell immunogenicity of PenG conjugated to foreign proteins (HEL or HSA), we next tested a native and otherwise tolerogenic self-antigen, mouse serum albumin (MSA), as a more physiologically relevant model 34 . First, pure MSA was pre-complexed with PenG in buffer under the optimised conditions, as described previously, and the resultant MSA-PenG was analysed via LC-MS/MS (Fig  S4 ). Animals were immunised three times (wk 0, 4 and 8) with or without alum and terminal (wk 10) IgG antibody titres were evaluated by ELISA against OVA-PenG (Fig.  2a ). 5/6 (83%) animals immunised with MSA-PenG in alum elicited a detectable serum endpoint titre against the adduct (median titre of ~ 5.2 × 10 3 ) and 3/8 (38%) responders in those immunised with MSA-PenG even without adjuvant (Fig.  2b ). Surprisingly, these data show that extrinsic adjuvantation is not required to elicit anti-adduct IgG responses even when presented on a modified self-protein. By contrast, no animal immunised with unmodified MSA alone with or without alum gave a detectable PenG-specific response ( P  = 0.033, Kruskal-Wallis test) (Fig.  2b ). Although the MSA-PenG-elicited PenG-specific antibody titres were lower compared with HSA-PenG (Figs.  1a , 2b ), these data show that a penicillin-specific IgG response can be generated even when using an otherwise highly T-immunorecessive ‘self-derived’ backbone, and that self-proteins can act as ‘non-self’ immunogenic carriers when modified by drug. No reactivity was detected against OVA in any group (Fig.  2c ).

a Sex-matched C57BL/6 mice were immunized three times and the serological response at the terminal timepoint (wk 10) was evaluated. b , c Mice were immunized three times (wk 0, 4 and 8) and the terminal (wk 10) IgG PenG-specific EPTs were evaluated. d Mice were bled and 2 mg/mL PenG was added and mixed end-to-end overnight. Animals were subsequently immunized (wk 0 and 4) intravenously with seum-PenG with or without alum. e , f Terminal IgG EPTs were evaluated. b , c , e , f Dots represent data from a single animal ( n  = 4–8 per group), and bars/text denotes the median (ND = not detected). Data were compared via a Kruskal–Wallis test (two-sided). g PenG-Ben structure. h Mice were given PenG-Ben intramuscularly. i PenG-specific IgG titres were evaluated. Dots represent data from a single animal ( n  = 27). Data were compared to pre-administration, evaluating the ratio of responders via Chi-squared test. a , d , h Created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license. Source data are provided as a Source Data file.

Finally, to develop an ex vivo model for self-protein:PenG adduct formation and immunogenicity, we tested the ability of complete mouse serum itself to act as a carrier for PenG. PenG-naïve mice were bled, serum isolated and incubated ex vivo with 2 mg/mL PenG for 16 h. This concentration was chosen to mimic a typical serum concentration in humans 35 . The resulting serum-adduct mixture was then administered intravenously (i.v.), into the autologous mice, with and without alum, to mimic a clinical route of penicillin administration (Fig.  2d ). Terminal IgG endpoint titres were evaluated against OVA-PenG, which revealed 6/8 (75%) of animals primed and boosted with adjuvanted serum-PenG responded with a detectable IgG titre against OVA-PenG and a median titre of ~ 2.5 × 10 3 (Fig.  2e,f ). Interestingly, only a single animal that received unadjuvanted serum-PenG gave a detectable PenG-specific IgG titre, implying that adjuvantation, such as might be generated by bacterial infection during therapeutic penicillin use, is likely required under such conditions to overcome a threshold for immunogenicity.

Immunogenicity of free penicillin via varying administration routes

Having shown that PenG is immunogenic when conjugated to diverse protein carriers including mouse serum, we next evaluated whether free penicillin, as would be administered in the clinic, might be sufficient to induce an antigen-specific antibody response. First, we tested whether PenG delivered i.v. daily to mice in 2 × 1 week-long courses was immunogenic (Fig  S5a ). However, no IgG or IgM drug-specific responses were detected when compared to mice given control PBS (Fig  S5b, c ). Second, the immunogenicity of orally administered antibiotic was evaluated, using the gut-stable oxo-homologue penicillin V (PenV). Unlike PenG, PenV is used for oral administration as it does not degrade under the acidic conditions of the stomach 36 . Mice were given PenV ad libitum for two 3.5-day intervals. Some mice were additionally given an i.p. dose of lipopolysaccharide (LPS) (0.5 mg/kg) to mimic possible systemic increase in endotoxin expected from a bacterial infection, where antibiotics such as PenG/V would be clinically used (Fig  S5d ). Despite this, no specific PenG titres were observed (Fig  S5e ). Notably, administration of LPS increased the background reactivity of serum: mice given PenG and LPS or drug-free water and LPS-only both exhibited modest reactivity against HSA-PenG, which we attribute to induction of B cells producing polyspecific IgG responses 37 .

Free drugs can be rapidly cleared; for instance, mice have an extremely high cardiac output 38 —renal clearance of PenG is efficient 39 , with a previously reported half-life of approximately 15 min. Fast clearance will restrict in vivo PenG adduct formation, ultimately reducing the probability for antigen-B cell encounter and BCR cross-linking. Therefore, to extend the availability of free drug in vivo, penicillin G benzathine (PenG-Ben) was used. This is a formulation of PenG as its dibenzylethylene diammonium salt that renders PenG effective for use in slow-release delivery (Fig.  2g ). When administered intramuscularly, PenG-Ben is solubilised over days to weeks, to release PenG systemically 40 . PenG-Ben was administered intramuscularly (i.m.) to 27 mice (Fig.  2h ). After the antibiotic course, 5/27 (18%) mice generated IgG responses to OVA-PenG, significantly higher than the proportion from pre-immune serum (χ 2  = 5.51, P  = 0.019) (Fig.  2i ). Reactivity against unmodified OVA was not detected (Fig  S5f ).

Cross-reactivity of anti-penicillin-adduct IgG responses are drug side-chain and core focused

PenG, of course, shares common structural homology with the other penicillins as well as other β-lactam antibiotics. We therefore screened the cross-reactivity of the HSA-PenG serological response. Ovalbumin (OVA) was modified with a set of penicillin antibiotics with differing side chains (Fig.  3 ), and with β-lactam antibiotics from other classes (including cephalosporins and carbapenems), using the previously determined conditions (Fig.  1c ). Extent of modification by drug was confirmed by evaluating reduced primary amine availability (Fig  S6 ). Autologous reactivity against OVA-PenG was the greatest of the diverse OVA-X panel tested, with a median IgG EPT of ~2.5 × 10 6 (Fig.  3a ). Interestingly, there was limited reactivity against OVA-ampicillin (median EPT of ~2.4 × 10 3 ), which differs only in a single benzylic amine substituent, and similarly carbenicillin (median EPT of ~ 4.9 × 10 3 ), which differs by its benzylic carboxyl substituent. However, there was considerable cross-reactivity against OVA-oxacillin (median EPT of ~ 8.5 × 10 4 ), despite the greater variation in sidechain compared with ampicillin. These data suggest that the polyclonal response tolerates some change in side-chain but that this may also be blocked by simple alterations at pivotal sites (such as the ampicillin H→NH 2 , or the carbenicillin H→COOH change). A subset of 1 st –4 th generation cephalosporin- and carbapenem-type antibiotics were also screened for cross-reactivity. HSA-PenG antisera displayed limited (albeit above the detection limit) cross-reactivity against these modified OVA antigens (IgG EPTs ~ 10 3 ). Since cephalosporins and carbapenems have differing β-lactam-encompassing cores 41 , these data suggested that the PenG-raised antibody response may be in part dependent on the 6-aminopenicillanic acid-derived core.

a HSA-PenG antisera were screened against a set of β-lactam antibiotic-modified OVA. Data reflect the IgG EPT against the drug adducts, with dots denoting reactivity from a single animal ( n  = 4), with bars/text denoting the median. b Sidechains of penicillins tested. Source data are provided as a Source Data file.

Clonotypic B cell responses to PenG adducts

To evaluate the B cell response at the clonal and molecular levels, PenG-specific B cells were isolated and variable regions cloned using techniques previously described 42 , 43 . Mice were immunised with HSA-PenG in alum and draining inguinal lymph nodes (iLN) were harvested 2 weeks post-prime (Fig.  4a ). To isolate PenG-specific B cells, requisite protein-based probes were synthesised by modifying another carrier protein (HIV-1 gp120) that we have validated as giving low background and high specificity in a other hapten-carrier contexts 29 . Gp120 was modified with PenG and then modified with fluorophore Alexa Fluor 647 using a corresponding NHS ester. We further tetramerised biotinylated gp120 with streptavidin-phycoerythrin. These antigen-displaying probes were then used to sort the PenG-specific B cells on pre-gated non-naïve (DUMP - B220 + IgD - ) B cells (Fig.  4b ; Fig  S7a ). B cells were sorted from four mice and cell clonality was inferred according to the VH sequences (Fig.  4c ).

a Immunisation schedule. Created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license. b Antigen probe sorting strategy on pre-gated non-naïve B cells (DUMP - B220+IgD - ). c Inferred clonal families from the PenG probe-sorted B cells (coloured) and singletons (grey). d Inferred V H gene segment across the sequenced B cells. Bar segment sizes proportional to the number of B cells of the same V H origin. Joining of the segments connote the shared utilisation between mice. Segment colours reflect the IGHV subgroup. e Logo plots of the CDRH1 and CDRH2 amino acid sequences from all sequenced B cells in all animals. f Fab binding specificity assay setup. g ELISA trace for a subset of 28 Fabs and ( h ) their OD1.25 intersection (estimate for EC 50 ) values. BAR-1 (grey) is an unrelated negative control specific to sialyllactose. Dots represent data from a single Fab clone ( n  = 28). Source data are provided as a Source Data file.

Considerable sharing of V H gene segments was observed between the mice, suggesting that similar clonotypes were raised across animals (Fig.  4d ; Fig  S7b ). Notably, the V H gene segments utilised were from four highly phylogenetically related subgroups: IGHV2 , IGHV5 , IGHV10 and IGHV14 . This striking homology is reflected in conservation of the CDRH1 and CDRH2 amino acid sequences across the mice (Fig.  4e ). Together, these data suggest that there are preferred structural and functional motifs encoded in these germline segments that facilitate binding with the drug. Immunogenetic analysis revealed a defined and ordered CDRH3; the length was bimodal, either short (5–6 aa) or long (17 aa) (Fig  S8a ). Short CDRH3s were dominated by an ARG motif for the first three residues, with a diverse C-terminal end, while the long class possesses a negative N-terminus, a neutral centre and a positive C-terminus (Fig  S8b ).

Clonal families were evaluated. The two largest families were isolated from mouse 1 and their germinal centre trees determined (Fig  S9 ). The largest clonal family (Fig  S6a ) exhibits a ‘clonal burst’ following the acquisition of the S11T mutation, a characteristic phenomenon reportedly associated with the acquisition of an affinity-improving mutation that renders the clone more competitive for antigen uptake and T cell help 44 . This mutation was also found in a separate clade on the same tree. The T85S mutation was also identified twice on separate clades.

Finally, to validate the PenG specificity of the antibody response, a subset of V-region pairs from all mice and with diverse gene segment origins were cloned, and corresponding fragment antigen-binding regions (FAbs) were expressed and purified. Binding was validated via ELISA; all FAbs bound the PenG adduct probe, while an unrelated antibody FAb (BAR-1) did not detectably bind (Fig.  4f–h ). These data confirmed that the sorting approach was highly specific.

Structural, biochemical and biophysical characterisation of the antibody response to PenG

We selected a subset of PenG-specific clones with divergent, representative CDRH3 lengths to further dissect the binding characteristics of the clonotypical response to penicilloyl adducts: MIL-1 ( IGHV5-6*01 ; 17 a.a. CDRH3), MIL-2 ( IGHV5-17*01 ; 6 a.a. CDRH3) and MIL-3 ( IGHV10-3*01 ; 9 a.a. CDRH3). We designed a soluble Lys-PenG ligand as a reductionist adduct mimic reflecting β-lactam display on the ε-amine of lysine residues (Fig.  5a ).

a Annotated chemical structure of PenG-lys construct used for uSTA analysis. b Saturation transfer difference spectrum was generated from the difference between the raw off-resonance (gaussian at 37 ppm) and the raw on-resonance (gaussian at 9 ppm) spectra. Data showing differences in para engagement in the benzene ring of MIL-2 fab versus MIL-1. c i. Heatmaps corresponding to saturation transfer efficiency of PenG-lys (1 mM) with MIL-1–3 (5 µM). ii. Histographic saturation transfer efficiencies. Red indicates high transfer efficiency. d SPR chip design. e Biophysical characterization of MIL-3 via SPR. f i. Top view of the x-ray structure of the MIL-3 Fab bound to the PenG-Lys (beige sticks). Both heavy (monochrome plum) and light chain (monochrome blue) CDRs are marked. Key residues within 4.0 Å of ligand aspects ii. phylacetamide, iii. thiazolidine and iv. lysine are shown. Hydrogen bonds are shown as black broken lines. Water is marked in red. Source data are provided as a Source Data file.

First, atomic resolution of the binding pose was dissected by protein NMR using universal saturation transfer analysis (uSTA) 29 , 45 . High transfer efficiencies were observed on the phenyl ring of Lys-PenG with all three FAb fragments, with the p -proton showing the greatest engagement for FAbs MIL-1 and MIL-3 (Fig.  5b,c ), suggestive of an end-on binding mode for the PenG phenylacetamide sidechain. Notably, the Lys residue itself exhibited minimal transfer efficiency in all cases, confirming, as was implied by the immunological data, that binding is dominated by drug adduct rather than peptide binding. Interestingly, the thiazolidine ring showed no significant engagement of the protons that are detected by uSTA. This engagement was further confirmed by drug adduct core binding (although less than drug sidechain) through observed transfer efficiencies of the proton at the stereogenic centre of the opened β -lactam ring (NC H ) as well as those of the benzylic sidechain CH 2 . These data revealed substantial uniformity in the binding poses adopted by the MIL series of antibodies.

Next, we determined the X-ray crystal structure of the PenG-Lys•MIL-3 complex at 2.2 Å (Table  S1 ). Three FAb molecules were present in the asymmetric unit (H (heavy)/L (light), A/B and C/D). In both H/L and A/B molecules, the phenylacetamide PenG sidechain and the thiazolidine moieties are well-defined in the electron density whilst the Cβ, Cδ and Cγ portion of the lysine moiety has weak density, indicating it is less well ordered and consistent with our observations by protein NMR (see above). The electron density is considerably weaker in A/B than in H/L most likely due to crystal contacts present in H/L (Fig  S10 a, Fig  S11 ). Apart from this region, the interactions between the ligand and the protein are conserved in both FAbs. In the third FAb molecule (C/D), CDRH loops are highly disordered, and no ligand was fitted.

Our analysis focuses on the H/L molecules (Fig.  5f ). The PenG benzene group is deeply buried in a narrow hydrophobic pocket sandwiched on one face of the benzene ring by CDR L loops (Tyr68 L , Tyr110 L ) and the β-turn formed by CRDH3, mainly the side chain of Ile120 H with some involvement of the main chain of Phe124 H , on the other face (Fig.  5f ii ., Fig S10b ). The interaction with Tyr68 L has strong π-stacking character whilst the planes of the rings benzene and Tyr110 L are offset 70° and thus hydrophobic in character. The methyl group of Cγ2 Ile120 H sits centred above plane of the ring.

The static X-ray structure does not disclose a simple explanation for the observation from NMR for the interactions of the p -hydrogen yet provides detail on PenG sidechain recognition. The nitrogen of the phenylacetamide forms a hydrogen bond with the carbonyl of Thr121 H whilst the phenylacetamide carbonyl oxygen bridges via a water molecule (W1) to the side chain hydroxyl of Tyr110 L . The β-turn conformation of CDRH3 is stabilised by hydrogen bonds between residues Thr122 H and Arg123 H and Tyr 51 H and Asp75 L .The carboxylate of the thiazolidine moiety makes a bidentate salt bridge to the guanidine group of Arg69 H and potentially a salt bridge with His54 H (Fig.  5f iii .); this binding mode is likely to powerfully contribute to binding enthalpy yet places the protons that are observable by NMR more remotely. The dimethyl group and thiazolidine ring make van der Waal contacts with Tyr115 L and Ile120 H . The nitrogen of the thiazolidine ring interacts with a highly coordinated water molecule (W3) bridging CDRs H1, H3 and the lysine linker. The main-chain mimic of the Lys makes five hydrogen bonds with the 3 10 helix of CDRH2 in H/L (Fig.  5f iv .) but only one in A/B (Fig  S10c ), consistent with differences in ordering and previously noted weaker interactions (see above). Torsion angle modification of the lysine side chain would permit conjugated protein to remain outside the binding pocket without any perturbation of the drug adduct thiazolidine and phenylacetamide interactions.

Finally, surface plasmon resonance (SPR) of FAb MIL-3 with a PenG-adduct chip revealed a K D value = 5.3 µM (90% CI: 3.996–7.37, Fig.  5d,e ).

Antibody clones are unlikely to sequester antibiotic during relevant treatments

The biophysical properties of the antibodies that we determined, the structural biology and the physiological concentrations of antibody in vivo all suggested that anti-PenG-adduct antibodies are unlikely to display any drug-sequestering effects. Nevertheless, this was evaluated by developing a model system to test whether PenG-specific antibodies could reduce the bacterial growth-inhibiting function of PenG. Whole antiserum was added to a culture of attenuated, unencapsulated Streptococcus pneumoniae 46 , chosen because of its high sensitivity to PenG (minimum inhibitory concentration (MIC) = 0.01 µg/mL). We conducted two assays: first, we evaluated the effects of anti-HSA-PenG antiserum on PenG sequestration in culture and, second, contrasted PenG kill zones when drug was pre-incubated with high quantities of purified recombinant control or anti-PenG MIL antibodies (Fig  S12 ). As expected, neither antiserum nor recombinant MIL antibody significantly inhibited antibiotic action.

The elicitation of ADA such as those against PenG may have broad implications, including the mediation of drug hypersensitivity 11 , 18 , 47 , 48 . The complex relationship between chemical reactivity and instability, pharmacokinetics, immunological factors and possible downstream functional effects of the resulting humoral responses are poorly understood. To address this, we have systematically evaluated the way in which the protein conjugation reactivity of the common β-lactam antibiotic, PenG, drives formation of antigenic complexes in vitro, and have established that diverse protein carriers are sufficient to propagate a drug-specific IgG response. Using a murine model, we characterised both the clonal B cell and antibody responses revealing striking clonal restriction and strong conservation of antibody-drug binding features despite diversity in CDRH3 length. Our data demonstrate that the production of PenG-specific antibodies is regulated at two distinct levels: (1) the covalent formation of protein adducts via lysine-amide formation is influenced by reaction conditions and time, and (2) immune engagement, including innate recruitment—such as via adjuvant which in vivo would be mimicked by bacterially-elicited inflammation—and T cell help. These factors ultimately determine the overall probability and magnitude of a downstream B cell response against PenG. Collectively, our data now provide a model for how adduct formation and immune engagement trigger the humoral response, phenomena that may inform both the study of allergy and potentially provide a rational approach to predict and even mitigate anti-penicillin antibody responses.

In a human context, most patients undergoing standard courses of PenG to treat bacterial infection exhibit anti-penicilloyl IgG antibodies thereafter 10 , 49 , 50 . These data appear initially incongruent with our murine data since animals given formulations of free penicillin either intravenously or in drinking water failed to exhibit a specific response. However, it is noteworthy that these selective effects may be attributed to dosing; humans are given as much as 50 mg/kg every 4–6 h of PenG via an intravenous line 51 , 52 . Moreover, higher murine cardiac output and drug clearance rate results in shorter drug half-life 38 , 39 . A critical role for PenG concentration, which will be higher in human circulation, in driving relevant adduct formation is supported by the serological data gathered from mice given both static pre-formed serum-PenG immunisations and intramuscular slow-release PenG-Ben. We did not titrate the downstream differences in murine versus human adaptive immune responses to penicillin, although differences are reported in the literature, particularly T cell-mediated responses to drug adducts 14 .

These experiments were not designed to explicitly evaluate allergic outcomes in mice. Isotype-switching to IgE and the downstream mast cell-mediated and other modes of allergic reactivity are determined by both genetic factors in the form of atopic predisposition to IgE production, and environmental factors 53 , 54 . However, the functional effect of the antibody response can not only be a feature of the Fc effector function or isotype—for example, whether type I versus type II hypersensitivity is imparted—but also the inherent binding mode and biophysical properties of the antibody. Our data now show that the murine B cell repertoire responds to the drug adduct with a dominant clonotypic family that primarily engages the side-chain constituent, phenylacetamide, and the carboxylate group of the thiazolidine, as evidenced by our complementary uSTA 45 and x-ray structural characterisation. These data point to a striking homogeneity of response, implying that B cell receptors and subsequent antibodies may be restricted in the binding solutions that they can adopt to recognise such a small non-protein antigen. This is consistent with previous immunological mapping studies of murine 55 , 56 , rabbit 27 and human 57 antibody responses in which side-chain reactivity appears prominent. Importantly, the biophysical features of these antibodies, with monovalent FAb K D in the low–mid µM range and a relatively fast k off , coupled with relatively low concentrations in vivo, are likely to discount even partial drug inhibitory effects.

Our approach, applied here to an archetypal inhibitor with a covalent bond-forming mode-of-action, now creates a potential roadmap for understanding the chemical, pharmacological and immunological factors governing whether a B cell response can be mounted against other protein-reactive covalent therapeutics, for example, the amide-forming acylating drug aspirin 58 . Having identified a dominant clonotype against the PenG adduct, this excitingly suggests a workflow that could be used as a model to ‘reverse engineer’ a PenG analogue that fails to engage the MIL clonotype, potentially reflecting a low/no allergenic alternative against a murine germline and that now instructs a proof of principle for germline-informed drug design in humans.

Ethics and permissions

All experiments were conducted under approved licenses and protocols, consistent with national (as authorised by the Home Office of the United Kingdom) regulations and University of Oxford guidelines.

Ex vivo modification of protein with β-lactam antibiotics

Carrier proteins (HEL, OVA, HSA, BSA and MSA) were purchased commercially (Merck) and dissolved in 0.1 M NaCO 3 (pH = 8, unless otherwise indicated) and concentration was adjusted to 1 mg/mL. β-lactam was added to a molar ratio of 1:200 per lysine residue of carrier protein, resulting in protein concentrations for HEL and HSA of 6.8 and 60 mg/mL respectively. The mixture was rotated end-to-end at 25 °C overnight and dialysed into PBS.

Denaturing MS

Reversed-phase chromatography was performed in-line prior to mass spectrometry using an Agilent 1100 HPLC system (Agilent Technologies inc.—Palo Alto, CA, USA). Concentrated protein samples were diluted to 0.02 mg/ml in 0.1% formic acid and 50 µl was injected on to a 2.1 mm ×12.5 mm Zorbax 5um 300SB-C3 guard column housed in a column oven set at 40 °C. The solvent system used consisted of 0.1% formic acid in ultra-high purity water (Millipore) (solvent A) and 0.1 % formic acid in methanol (LC-MS grade, Chromasolve) (solvent B). Chromatography was performed as follows: Initial conditions were 90% A and 10% B and a flow rate of 1.0 mL/min. After 15 s at 10% B, a two-stage linear gradient from 10% B to 80% B was applied, over 45 s and then from 80% B to 95% B over 3 s. Elution then proceeded isocratically at 95% B for 1 mins 12 s followed by equilibration at initial conditions for a further 45 s. Protein intact mass was determined using a 1969 MSD-ToF electrospray ionisation orthogonal time-of-flight mass spectrometer (Agilent Technologies Inc.—Palo Alto, CA, USA). The instrument was configured with the standard ESI source and operated in positive ion mode. The ion source was operated with the capillary voltage at 4000 V, nebulizer pressure at 60 psig, drying gas at 350 °C and drying gas flow rate at 12 L/min. The instrument ion optic voltages were as follows: fragmentor 250 V, skimmer 60 V and octopole RF 250 V. Obtained MS spectra were processed and deconvoluted using the Agilent MassHunter Qualitative Analysis (B.07.00) software.

Approximately 5 µg protein was reduced, loaded and run on an SDS-PAGE. Gel bands were excised and washed sequentially with HPLC grade water followed by 1:1 (v/v) MeCN/H 2 O. Gel bands were dried (via vacuum centrifuge), treated with 10 mM dithiothreitol (DTT) in 100 mM NH 4 HCO 3 and incubated for 45 min at 56 °C with shaking. DTT was removed and 55 mM iodoacetamide (in 100 mM NH4HCO3) was added and incubated for 30 min in the dark. All liquid was removed and gels were washed with 100 mM NH 4 HCO 3 /MeCN as above. Gels were dried and 12.5 ng/µL trypsin was added separately and incubated overnight at 37 °C. Samples were then washed and peptides were extracted and pooled with sequential washes with 5% (v/v) formic acid (FA) in H 2 O and MeCN. Dried samples were reconstituted in 2% MeCN 0.05% trifluoroacetic acid and run by LC-MS.

Samples were analysed using an Ultimate 3000 UHPLC coupled to an Orbitrap Q Exactive mass spectrometer (Thermo Fisher Scientific). Peptides were loaded onto a 75 µm × 2 cm pre-column and separated on a 75 µm × 15 cm Pepmap C18 analytical column (Thermo Fisher Scientific). Buffer A was 0.1% FA in H 2 O and buffer B was 0.1% FA in 80% MeCN with 20% H 2 O. A 40-min linear gradient (0% to 40% buffer B) was used. A universal HCD identification method was used. Data was collected in data-dependent acquisition mode with a mass range 375 to 1500 m/z and at a resolution of 70,000. For MS/MS scans, stepped HCD normalized energy was set to 27, 30 and 33% with orbitrap detection at a resolution of 35,000.

Raw data was first searched using the FragPipe (v19.1) 59 Open Search pipeline to determine the modified mass shift caused by PenG conjugation. Approximately 2.3% of peptide spectral matches (PSM) had an unannotated mass shift of 334.099 Da. Next, to determine site specificity and occupancy, raw data was searched using Proteome Discoverer (3.0.0.757). In-house curated FASTA databases were used. The digestion enzyme was set to trypsin with a maximum of 2 miss cleavages. A 10 ppm precursor mass tolerance and 0.6 Da fragment mass tolerance were allowed. Oxidation (+15.995 Da) of methionine and PenG conjugation (+334.099 Da) of lysines and protein N-termini were set to dynamic modifications. Carbamidomethylation (+57.021 Da) of cysteines was set as a static modification. Target false discovery rate (FDR) for peptide spectrum matches, peptide and protein identification was set to 1%. To approximate site-specific occupancy of PenG conjugation the total number of PSMs of peptides containing a specific lysine site in a PenG modified state were expressed as a percentage compared to the total number of PSMs containing the given lysine in both the modified and unmodified state.

To estimate lysine reactivity in HSA, raw data was searched against a database of known contaminants which contains the canonical HSA sequence using FragPipe (21.1) using a standard closed search parameters with an additional variable modification confined to lysines (+334.099). IonQuant (as implemented in FragPipe) was used to calculate peptide intensities with default parameters. Intensity of each modification-specific peptide was normalised against total intensity of all HSA peptides in treated and control samples. For each position, a ratio of intensity of unmodified peptide (without missed cleavages, as present in control sample) between the treated sample and the control was calculated to estimate the extent of lysine modification at that position after treatment.

LC-MS/MS raw data and search results have been deposited to the ProteomeXchange Consortium ( http://proteomecentral.proteomexchange.org ) via the PRIDE partner repository 60 with the dataset identifier: PXD052026.

Free primary amine ELISA

After the drugs were conjugated to a carrier protein using the method outlined above, the relative abundance of free amines was assessed to determine the extent of lysine modification. Protein samples (5 μg) were dissolved in 10 μL of PBS and mixed with 40 μL of 0.1 M sodium bicarbonate buffer. A 5% solution of 2,4,6-trinitrobenzenesulfonic acid (TNBSA) was diluted at a ratio of 1:500 in the bicarbonate buffer, and 25 μL of this mixture was added to the protein samples. Following a 2 h incubation period at 37 °C, 25 μL of 10% SDS and 12.5 μL of 1 M HCl were added. The absorbance at 335 nm was then measured.

Mice and immunisation formulations

Wild-type, specific pathogen-free, sex-matched, 6–8-week-old C57BL/6 mice were purchased from Charles River. Animals were monitored daily and were provided standard chow and water ad libitum . Immunisation formulations and schedules are outlined in the results. Mice were bled periodically from the tail vein. Animals were sacrificed via a rising CO 2 gradient and subsequent cervical dislocation schedule 1 procedure.

Serum samples were serially diluted and transferred onto an antigen-coated and blocked SpectraPlate-96 (PerkinElmer) plate. Binding was detected with an anti-mouse IgG-HRP (STAR120P, Bio-Rad). ELISAs were developed using 1-Step-Ultra TMB ELISA substrate (Life Technologies), terminating the reaction with 0.5 M H 2 SO 4 . For competition ELISAs, serial dilution of soluble ligands as preincubated with antiserum at the pre-determined EC 50 concentration for 1 h. The antisera and ligand mixtures were subsequently transferred onto the antigen-coated and blocked plates and ELISA conducted, as previously outlined. Cytokine ELISAs were performed using commercially available kits (Life Technologies), screening supernatant from antigen-restimulated splenocytes.

Optical densities were measured at 450 and 570 nm on a Spectramax M5 plate reader (Molecular Devices). After background subtraction, logistic dose-response curves were fitted in GraphPad Prism. Endpoint titres were determined as the point at which the best-fit curve reached an OD 450-570 value of 0.01, a value which was always > 2 standard deviations above background.

B cell sorting

Penicilloyl-specific B cells were isolated using antigen probes. Gp120-PenG was modified with an NHS-esterified AF647 dye, as per the manufacturer’s instructions (Life Technologies). To improve true antigen-specific cell sorting efficiency, a negative backbone-specific probe was synthesised, wherein biotinylated gp120 was tetramerised with PE-conjugated strepdavidin (Biolegend).

Single cell suspensions were stained with LIVE/DEAD Fixable Blue and Fc receptors blocked. Surface staining was performed using anti-mouse F4/80-PE (1:200, BM8, Biolegend), anti-mouse Gr-1 (1:200, RB6-8C5, Biolegend), anti-mouse CD3-PE (1:200, 17A2, Biolegend), anti-mouse CD4-PE (1:200, RM4-5, Biolegend), anti-mouse CD8-PE (1:200, RPA-T8, Biolegend), anti-mouse B220-eFluor450 (1:100, RA3-6B2, BD Biosciences), anti-mouse IgD-AF700 (1:200, 11-26 c.2a, Biolegend), anti-mouse IgM-PE/Cy7 (1:200, R6-60.2, BD Biosciences), anti-mouse IgG1-FITC (1:200, A85-1, BD Biosciences), anti-mouse IgG2a/2b-FITC (1:200, R2-40, BD Bioscience) and antigen probes (10 μg/mL). Cells were stained on ice for 1 h, washed and stored on a BD FACSAriaFusion (BD Biosciences). Single cells were sorted into MicroAmp Optical 96-well PCR plates (Life Technologies), isolating LIVE/DEAD - DUMP - B220 + IgD - gp120 - gp120-PenG + events. Cells were sorted directly into 5 µL if 1X TCK buffer supplemented with 1% 2-ME and stored at −80 °C until use.

Thymidine incorporation

Whole splenocytes were stimulated in vitro with 10 µg/mL antigen in cRMPI for 16 h in a flat-bottom 96-well plate. During the final 6 h stimulation, each well was spiked with 0.037 mBq tritiated thymidine (Perkin Elmer). Cells were transferred and lysed on glass filter mats (Perkin Elmer) using a Micro 96 Harvester (Skatron Instruments). Tritium incorporation was measured using Betaplate Scint and a Microbeta Trilux Scintilation counter (Perkin Elmer).

Intracellular cytokine staining

Whole splenocytes were stimulated in vitro with 10 µg/mL antigen in cRPMI for 16 h. For the final 6 h, 5 µg/mL brefeldin A (Biolegend) was added to suspend ET–Golgi trafficking and block cytokine secretion. Cells were stained with TruStain mouse FcX Plus (Biolegend) and LIVE/DEAD Fixable Blue in PBS with 2 mM EDTA for 30 mins. Surface markers were subsequently stained: PE-conjugated anti-mouse CD3 (dilution: 1:200, clone: 17A2, manufacturer: Biolegend), APC-conjugated anti-mouse CD4 (1:200, RM4-5, Biolegend), AF700-conjugated anti-mouse CD8 (1:200, RPA-T8, Biolegend). Following fixation and permeabilization (Biolegend), cells were stained with PE/DAZZLE-conjugated anti-mouse IFN- γ (1:100, XMG1.2, Biolegend). Cells were washed and data was acquired on the BD Fortessa X-20 (BD Biosciences), collecting 500,000 events per sample.

Variable region cloning and antibody expression

B cell receptor variable regions were recovered, as previously described 29 . Briefly, RNA was captured on RNAClean XP beads (Beckman Coulter) and washed with 70% ethanol. RNA was eluted and cDNA was synthesised using SuperScript III (Life Technologies) with random primers (Life Technologies). VH and VK regions were recovered 43 and Q5 polymerase (New England Bioscience), sequencing the amplicons via Sanger. VH amplicon sequences were used to determine B cell clonality.

To validate that the sequences were specific to the penicilloyl adducts, antibodies were recombinantly expressed. The VH/VK amplicons were incorporated into expression vectors: vector-overlapping adapters were incorporated via PCR 42 , and the V regions were inserted into pre-cut recombinant FAb expression 44 vector via Gibson reaction (New England Bioscience). Vector products were transiently transfected into HEK 293Freestyle cells and FAb was purified from supernatant using Ni-NTA resin.

Immunogenetic analyses

Analyses were performed using VH regions. Sequences were aligned to the murine reference genome using the Immunogenetics Information System (IMGT; https://www.imgt.org/IMGT_vquest/input ), as described previously 29 . Sequence outputs of poor quality or those unproductive were excluded from our analyses. Alignments of CDRs were visualised using WebLogo 61 . Clonal lineages were evaluated using GCTree 62 .

SPR was performed using a Biacore T200 instrument. Details of chip design, synthesis and testing, refer to Document  S1 . FAb binding was evaluated by sequentially injecting serial dilutions at a flow rate of 10 µL/min.

Samples for uSTA were prepared by buffer-exchange of purified FAb fragments with D 2 O PBS using Amicon 30 K MWCO. All NMR experiments were conducted on Bruker Avance Neo 600 MHz spectrometer at 25 o C equipped with QCIF cryoprobe and a SampleJet, running TopSpin 4.2.0. The uSTA experiments were recorded using a pseudo-3D pulse sequence based on stddiffesgp.2 from the standard Bruker library as described previously 45 . The following saturation times were used: 0.1, 0.3, 0.5, 0.7, 0.9, 1.1, 1.3, 1.5, 1.7, 1.9, 2, 2.5, 3, 3.5, 4 and 5 s. Low-power gaussian excitation pulse was applied at 9 ppm and 37 ppm for the on- and off-resonance spectra respectively, where the specific choice of excitation at 9 ppm minimised artefacts in a ligand-only spectrum 45 (Fig.  S1 ). All experiments were recorded with 16 scans per transient, 32768 complex points and sweep width of 16.01 ppm for a total acquisition time of 7 h 41 min. These were acquired on a protein-only, a ligand-only and a mixed protein/ligand sample. Data were processed using the nmrPipe module embedded within the uSTA workflow, where the protein-only data is subtracted from the mixture to give a uSTA transfer efficiently, from which values from the ‘ligand only’ sample are then subtracted to account for residual ligand excitation, as previously reported 45 . The ‘heatmaps’ were generated by mapping the uSTA transfer efficiency from the detectable protons onto heteroatoms (carbon, sulphur, oxygen and nitrogen) using a 1/r 6 dependency, before rendering in pymol as described previously 45 .

X-ray crystallography

MIL-3 FAb was loaded onto a gel filtration Superdex 200 column 10/30 (GE Healthcare) in 10 mM Tris-HCl, pH 7.5, 150 mM NaCl. Co-crystals appeared at 20 C after a week from a hanging drop of 0.1 μL of protein solution (15 mg/mL with 2.5 mM PenG-Lys with 0.1 μL of reservoir solution containing 20% (w/v) PEG 6000, 0.1 M MES pH 6, 0.2 M ammonium chloride in vapor diffusion with reservoir. Crystals were frozen with the same solution containing 16% glycerol. Data were collected at the Diamond light source oxfordshire (beamlines I04). Data were processed with XIA2 63 , 64 , 65 , 66 . Structure has been solved by molecular replacement using PHASER and pdb file 7bh8 for VL, CH and CL domains and 7n09 for VH domain. The structure was builded with Autobuild program, refined with REFINE of PHENIX with NCS restraints 67 and adjusted with COOT 68 . Coordinates and topologies of ligands were generated by AceDRG 69 .

Microbiological assays

The attenuated, unencapsulated lab strain Streptococcus pneumoniae D39 (Δ cps2A’ -Δ cps2H’ ) 46 was routinely grown in tryptic soy broth (TSB) (BD Biosciences) at 37 °C (standing incubation) in a 5% CO 2 atmosphere.

Microbroth dilutions of S . pneumoniae D39 revealed a PenG MIC of 0.01 μg/mL (that is, the lowest antibiotic concentration that prevented bacterial growth) (data not shown). For the liquid antibiotic rescue assay, 2 ng of PenG in 10 μL PBS were pre-incubated with 10 μL of antisera for 2 h in a flat-bottom 96-well plate. 180 μL of exponentially growing S. pneumoniae cells (OD 600  = 0.2) were added and incubated overnight (such that the final PenG concentration was 0.01 μg/mL). The following day, bacterial growth was measured using a Spectramax M5 plate reader (Molecular Devices), evaluating the OD 600 nm as a proxy for bacterial density.

For disk diffusion assays, 0.1 μg of PenG and mAb (1:50 molar ratio) were spotted onto paper disks. The disk was placed onto a blood agar plate (Merck), carrying a 5 ml nutrient soft agar overlay with 200 μL exponentially growing S. pneumoniae cells (OD 600  = 0.2). Plates were incubated overnight. The following day, the kill zone diameter was manually measured. PenG-only, PenG-raised antibody clones were tested, as well as an irrelevant HIV-1-specific antibodies were tested.

Data and statistics

Flow cytometry data was evaluated using FlowJo V.10.8.2 for Mac. Statistical analyses were conducted in either GraphPad Prism V.10.0.1 or in RStudio V.4.1. Statistical test details are provided in the results, figures and associated figure legends.

Data availability

Data reported in the manuscript are supplied as separate source data files or deposited as otherwise referred throughout or can alternatively be requested directly from the author. LC-MS/MS raw data and search results have been deposited to the ProteomeXchange Consortium ( http://proteomecentral.proteomexchange.org ) via the PRIDE partner repository 60 with the dataset identifier: PXD052026. The structure of BAR-1 bound to PenG-Lys is deposited in the protein database PDB ( https://www.rcsb.org ), under the accession number 8QXC. No custom code was developed for this manuscript. Reagents are available where applicable through an institutional MTA agreement.  Source data are provided with this paper.

Eisen, H. N. & Siskind, G. W. Variations in affinities of antibodies during the immune response * . Biochemistry 3 , 996–1008 (1964).

Article   CAS   PubMed   Google Scholar  

Mitchison, N. A. The carrier effect in the secondary response to hapten-protein conjugates. I. Measurement of the effect with transferred cells and objections to the local environment hypothesis. Eur. J. Immunol. 1 , 10–17 (1971).

Landsteiner, K. & Jacobs, J. Studies on the sensitization of animals with simple chemical compounds. Ii. J. Exp. Med 64 , 625–639 (1936).

Landsteiner, K. & van der Scheer, J. On the specificity of serological reactions with simple chemical compounds. J. Exp. Med 54 , 295–304 (1931).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Park, B. K., Naisbitt, D. J., Gordon, S. F., Kitteringham, N. R. & Pirmohamed, M. Metabolic activation in drug allergies. Toxicology 158 , 11–23 (2001).

Chen, C.-B. et al. An updated review of the molecular mechanisms in drug hypersensitivity. J. Immunol. Res. 2018 , 6431694 (2018).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Potashman, M. H. & Duggan, M. E. Covalent modifiers: an orthogonal approach to drug design. J. Med. Chem. 52 , 1231–1246 (2009).

Kim, H., Hwang, Y. S., Kim, M. & Park, S. B. Recent advances in the development of covalent inhibitors. RSC Med. Chem. 12 , 1037–1045 (2021).

Kostka, P., Barnett, W. H. & Kwan, C. Y. Identification and characterization of high-affinity Ca2(+)-ATPase associated with axonal plasma membranes of dog mesenteric nerves. Neurochem. Res. 15 , 833–841 (1990).

Lafaye, P. & Lapresle, C. Fixation of penicilloyl groups to albumin and appearance of anti-penicilloyl antibodies in penicillin-treated patients. J. Clin. Investig. 82 , 7–12 (1988).

Weltzien, H. U. & Padovan, E. Molecular features of penicillin allergy. J. Investig. Dermatol. 110 , 203–206 (1998).

Levine, B. B. & Ovary, Z. Studies on the mechanism of the formation of the penicillin antigen. J. Exp. Med. 114 , 875–940 (1961).

Dupnik, K. What’s “hapten”-ing in β-lactam hypersensitivity? Sci. Transl. Med. 9 , eaan3774 (2017).

Article   PubMed   Google Scholar  

Meng, X. et al. Definition of the nature and hapten threshold of the β-lactam antigen required for T cell activation in vitro and in patients. J. Immunol. 198 , 4217–4227 (2017).

Gill, F. A. The association of increased spontaneous lymphocyte transformation in vitro with clinical manifestations of drug hypersensitivity. J. Immunol. 98 , 778–785 (1967).

Blumenthal, K. G. et al. Peripheral blood eosinophilia and hypersensitivity reactions among patients receiving outpatient parenteral antibiotics. J. Allergy Clin. Immunol. 136 , 1288–1294.e1 (2015).

Brander, C. et al. Heterogeneous T cell responses to beta-lactam-modified self-structures are observed in penicillin-allergic individuals. J. Immunol. 155 , 2670–2678 (1995).

Fontaine, C. et al. Relevance of the determination of serum-specific IgE antibodies in the diagnosis of immediate?-lactam allergy. Allergy 62 , 47–52 (2007).

Neugut, A. I., Ghatak, A. T. & Miller, R. L. Anaphylaxis in the United States. Arch. Intern. Med. 161 , 15 (2001).

Shenoy, E. S., Macy, E., Rowe, T. & Blumenthal, K. G. Evaluation and management of penicillin allergy. JAMA 321 , 188 (2019).

Savic, L. et al. BSACIguideline for the set‐up of penicillin allergy de‐labelling services by non‐allergists working in a hospital setting. Clin. Exp. Allergy 52 , 1135–1141 (2022).

West, R. M. et al. Warning: allergic to penicillin’: association between penicillin allergy status in 2.3 million NHS general practice electronic health records, antibiotic prescribing and health outcomes. J. Antimicrobial Chemother. 74 , 2075–2082 (2019).

Article   CAS   Google Scholar  

Shenoy, E. S., Macy, E., Rowe, T. & Blumenthal, K. G. Evaluation and management of penicillin allergy: a review. JAMA 321 , 188–199 (2019).

Qiao, H.-L. et al. Hypersensitivity reactions to penicillins: studies in a group of patients with negative benzylpenicillin G skin test. J. Clin. Pharm. Ther. 34 , 249–254 (2009).

Qiao, H.-L., Gao, N., Jia, L.-J., Yang, J. & Tian, X. Specific IgG antibodies in sera in patients with penicillin allergy. Clin. Exp. Med. 9 , 105–111 (2009).

Dandliker, W. B., Halbert, S. P., Florin, M. C., Alonso, R. & Schapiro, H. C. Study of penicillin antibodies by fluorescence polarization and immunodiffusion. J. Exp. Med. 122 , 1029–1048 (1965).

Brandriss, M. W., Smith, J. W. & Steinman, H. G. Common antigenic determinants of penicillin g, cephalothin and 6-aminopenicillanic acid in rabbits. J. Immunol. 94 , 696–704 (1965).

Meng, X. et al. Direct evidence for the formation of diastereoisomeric benzylpenicilloyl haptens from benzylpenicillin and benzylpenicillenic acid in patients. J. Pharmacol. Exp. Therapeutics 338 , 841–849 (2011).

Deimel, L. P. et al. Engineered display of ganglioside-sugars on protein elicits a clonally and structurally constrained B cell response. bioRxiv https://doi.org/10.1101/2023.06.03.543556 (2023).

Peters, N. C., Hamilton, D. H. & Bretscher, P. A. Analysis of cytokine-producing Th cells from hen egg lysozyme-immunized mice reveals large numbers specific for ‘cryptic’ peptides and different repertoires among different Th populations. Eur. J. Immunol. 35 , 56–65 (2005).

Ariza, A. et al. Study of protein haptenation by amoxicillin through the use of a biotinylated antibiotic. PLoS ONE 9 , e90891 (2014).

Jenkins, R. E. et al. β-lactam antibiotics form distinct haptenic structures on albumin and activate drug-specific T-lymphocyte responses in multiallergic patients with cystic fibrosis. Chem. Res Toxicol. 26 , 963–975 (2013).

Whitaker, P. et al. Mass spectrometric characterization of circulating and functional antigens derived from piperacillin in patients with cystic fibrosis. J. Immunol. 187 , 200–211 (2011).

Wuttge, D. M., Bruzelius, M. & Stemme, S. T-cell recognition of lipid peroxidation products breaks tolerance to self proteins. Immunology 98 , 273–279 (1999).

Dagan, R. Evidence to support the rationale that bacterial eradication in respiratory tract infection is an important aim of antimicrobial therapy. J. Antimicrobial Chemother. 47 , 129–140 (2001).

Poschet, J. F., Hammond, S. M. & Fairclough, P. D. Characterisation of penicillin G uptake in human small intestinal brush border membrane vesicles. Gut 44 , 620 (1999).

Gronowicz, E. S., Doss, C. & Schröder, J. Activation to IgG secretion by lipopolysaccharide requires several proliferation cycles. J. Immunol. 123 , 2057–2062 (1979).

Kreissl, M. C. et al. Noninvasive measurement of cardiovascular function in mice with high-temporal-resolution small-animal PET. J. Nucl. Med. 47 , 974–980 (2006).

PubMed   Google Scholar  

Erlendsdottir, H. et al. Penicillin pharmacodynamics in four experimental pneumococcal infection models. Antimicrob. Agents Chemother. 45 , 1078–1085 (2001).

Merad, B. et al. Blood determination of benzathine-penicillin used in acute joint rheumatism prophylaxis. Arch. Inst. Pasteur Alger 62 , 32–39 (1998).

CAS   PubMed   Google Scholar  

Turner, J. et al. The chemical relationship among beta-lactam antibiotics and potential impacts on reactivity and decomposition. Front. Microbiol 13 , 807955 (2022).

Article   PubMed   PubMed Central   Google Scholar  

Viant, C., Escolano, A., Chen, S. T. & Nussenzweig, M. C. Sequencing, cloning, and antigen binding analysis of monoclonal antibodies isolated from single mouse B cells. STAR Protoc. 2 , 100389 (2021).

Ho, I. Y. et al. Refined protocol for generating monoclonal antibodies from single human and murine B cells. J. Immunol. Methods 438 , 67–70 (2016).

Tas, J. M. J. et al. Visualizing antibody affinity maturation in germinal centers. Science 351 , 1048–1054 (2016).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Buchanan, C. J. et al. Pathogen-sugar interactions revealed by universal saturation transfer analysis. Science 377 , eabm3125 (2022).

Lanie, J. A. et al. Genome sequence of Avery’s virulent serotype 2 strain D39 of Streptococcus pneumoniae and comparison with that of unencapsulated laboratory strain R6. J. Bacteriol. 189 , 38–51 (2007).

Kohler, H., Pashov, A. & Kieber-Emmons, T. The promise of anti-idiotype revisited. Front Immunol. 10 , 808 (2019).

Weiss, M. E. & Adkinson, N. F. Immediate hypersensitivity reactions to penicillin and related antibiotics. Clin. Exp. Allergy 18 , 515–540 (1988).

Gordon, S. M. et al. The response of symptomatic neurosyphilis to high-dose intravenous penicillin G in patients with human immunodeficiency virus infection. N. Engl. J. Med. 331 , 1469–1473 (1994).

Neftel, K. A., Wälti, M., Schulthess, H. K. & Gubler, J. Adverse reactions following intravenous penicillin-G relate to degradation of the drug in vitro. Klin. Wochenschr. 62 , 25–29 (1984).

Huynh, D. et al. Amoxicillin and penicillin G dosing in pediatric community‐acquired pneumococcal pneumonia in the era of conjugate pneumococcal vaccines. Pharmacotherapy https://doi.org/10.1002/phar.2756 (2023).

Metsvaht, T. et al. High variability in the dosing of commonly used antibiotics revealed by a Europe-wide point prevalence study: implications for research and dissemination. BMC Pediatr. 15 , 41 (2015).

Corry, D. B. & Kheradmand, F. Induction and regulation of the IgE response. Nature 402 , 18–23 (1999).

Article   ADS   Google Scholar  

Eckl-Dorna, J. et al. Tracing IgE-producing cells in allergic patients. Cells 8 , 994 (2019).

Mayorga, C. et al. Epitope mapping of beta-lactam antibiotics with the use of monoclonal antibodies. Toxicology 97 , 225–234 (1995).

de Haan, P., de Jonge, A. J., Verbrugge, T. & Boorsma, D. M. Three epitope-specific monoclonal antibodies against the hapten penicillin. Int Arch. Allergy Appl. Immunol. 76 , 42–46 (1985).

Harle, D. G. & Baldo, B. A. Identification of penicillin allergenic determinants that bind IgE antibodies in the sera of subjects with penicillin allergy. Mol. Immunol. 27 , 1063–1071 (1990).

Liyasova, M. S., Schopfer, L. M. & Lockridge, O. Reaction of human albumin with aspirin in vitro: Mass spectrometric identification of acetylated lysines 199, 402, 519, and 545. Biochem. Pharm. 79 , 784–791 (2010).

Kong, A. T., Leprevost, F. V., Avtonomov, D. M., Mellacheruvu, D. & Nesvizhskii, A. I. MSFragger: ultrafast and comprehensive peptide identification in mass spectrometry–based proteomics. Nat. Methods 14 , 513–520 (2017).

Vizcaíno, J. A. et al. ProteomeXchange provides globally coordinated proteomics data submission and dissemination. Nat. Biotechnol. 32 , 223–226 (2014).

Crooks, G. E., Hon, G., Chandonia, J.-M. & Brenner, S. E. WebLogo: a sequence logo generator: Fig. 1. Genome Res. 14 , 1188–1190 (2004).

DeWitt, W. S. III, Mesin, L., Victora, G. D., Minin, V. N. & Matsen, F. A. IV Using genotype abundance to improve phylogenetic inference. Mol. Biol. Evol. 35 , 1253–1265 (2018).

Winter, G. xia2: an expert system for macromolecular crystallography data reduction. J. Appl Crystallogr 43 , 186–190 (2010).

Article   ADS   CAS   Google Scholar  

Winter, G. et al. DIALS: implementation and evaluation of a new integration package. Acta Crystallogr. D: Struct. Biol. 74 , 85–97 (2018).

Article   ADS   CAS   PubMed   Google Scholar  

Beilsten-Edmands, J. et al. Scaling diffraction data in the DIALS software package: algorithms and new approaches for multi-crystal scaling. Acta Crystallogr. D: Struct. Biol. 76 , 385–399 (2020).

Winn, M. D. et al. Overview of the CCP 4 suite and current developments. Acta Crystallogr. D: Biol. Crystallogr. 67 , 235–242 (2011).

Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr. D: Struct. Biol. 75 , 861–877 (2019).

Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D: Biol. Crystallogr. 60 , 2126–2132 (2004).

Article   ADS   PubMed   Google Scholar  

Long, F. et al. AceDRG : a stereochemical description generator for ligands. Acta Crystallogr. D: Struct. Biol. 73 , 112–122 (2017).

Download references

Acknowledgements

The authors thank The Sir William Dunn School of Pathology flow cytometry facility, SPR facility and animal house staff. We extend our gratitude to the Rosetrees Trust, who have supported this work through the Interdisciplinary Award (ID2020/100023). We are also grateful for the funding provided by the Wellcome Trust (224212/Z/21/Z). Additionally, we thank the Wellcome Trust (grant ref: 095872/Z/10/Z) and the Engineering and Physical Sciences Research Council (grant ref: EP/R029849/1) for the instrumental upgrades of the 600-mHz and 950-MHz NMR spectrometers, as well as support from the University of Oxford Institutional Strategic Support Fund, the John Fell Fund, and the Edward Penley Abraham Cephalosporin Fund. A.J.B. is supported by ERC grant (101002859). For the purpose of Open Access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission. The Chemistry theme at the Rosalind Franklin Institute is sustained by the EPSRC (V011359/1 (P)). We would like to thank the Membrane Protein Laboratory at Diamond Light Source (funded by Wellcome Trust grant 223727/Z/21/Z) for help and support. C.M.K. is supported by an EPA Cephalosporin Junior Research Fellowship from Linacre College Oxford. L.P.D. is supported by the Clarendon Fund, and Q.J.S. is a Jenner Vaccine Institute Investigator and James Martin School Senior Fellow. We are grateful for the technical advice of Anton van der Merwe (The Sir William Dunn School of Pathology, University of Oxford) in the biophysical analyses.

Author information

Lachlan P. Deimel

Present address: Laboratory of Molecular Immunology, The Rockefeller University, New York, NY, 10065, USA

Authors and Affiliations

Sir William Dunn School of Pathology, University of Oxford, Oxford, OX1 3RE, UK

Lachlan P. Deimel, Mikhail Kutuzov, Carolin M. Kobras, Mathew Stracy & Quentin J. Sattentau

Rosalind Franklin Institute, Harwell Science and Innovation Campus, Oxford, OX11 0FA, UK

Lucile Moynié, Guoxuan Sun, Viliyana Lewis, Abigail Turner, Charles J. Buchanan, Yana Demyaneko, Shabaz Mohammed, Andrew J. Baldwin, James Naismith & Benjamin G. Davis

Department of Chemistry, University of Oxford, Oxford, OX1 3TA, UK

Charles J. Buchanan, Shabaz Mohammed, Andrew J. Baldwin & Benjamin G. Davis

Kavli Institute for Nanoscience Discovery, Dorothy Crowfoot Hodgkin Building, University of Oxford, Oxford, OX1 3QU, UK

Charles J. Buchanan, Sean A. Burnap, Weston B. Struwe & Andrew J. Baldwin

Department of Biochemistry, University of Oxford, Oxford, OX1 3QU, UK

Sean A. Burnap, Shabaz Mohammed & Weston B. Struwe

Department of Pharmacology, University of Oxford, Oxford, OX1 3QT, UK

Yana Demyaneko & Benjamin G. Davis

The Max Delbrück Centre for Molecular Medicine, Campus Berlin-Buch, 13125, Berlin, Germany

Quentin J. Sattentau

Experimental and Clinical Research Center (ECRC), Charité Universitätsmedizin Berlin and Max-Delbrück-Center for Molecular Medicine, Lindenberger Weg 80, 13125, Berlin, Germany

You can also search for this author in PubMed   Google Scholar

Contributions

Conceptualization of project: L.P.D., B.G.D., Q.J.S.; Methodology: L.P.D., L.M., G.S., V.L., A.T., C.J.B., S.A.B., M.K., C.M.K., M.S., W.B.S., A.J.B., J.N., B.G.D., Q.J.S.; Investigation: L.P.D., L.M., G.S., V.L., A.T., C.J.B., S.A.B., M.K., Y.D., C.M.K., W.B.S.; Funding acquisition: M.S., W.B.S., A.J.B., J.N., B.G.D., Q.J.S.; Project administration: M.S., W.B.S., A.J.B., J.N., B.G.D., Q.J.S.; Supervision: S.M., M.S., W.B.S., A.J.B., J.N., B.G.D., Q.J.S.; Writing—original draft: LPD.; Writing—review & editing: L.P.D., B.G.D., Q.J.S.

Corresponding authors

Correspondence to Lachlan P. Deimel , Benjamin G. Davis or Quentin J. Sattentau .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature Communications thanks Jeremy Derrick, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information, peer review file, source data, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Deimel, L.P., Moynié, L., Sun, G. et al. Covalent penicillin-protein conjugates elicit anti-drug antibodies that are clonally and functionally restricted. Nat Commun 15 , 6851 (2024). https://doi.org/10.1038/s41467-024-51138-7

Download citation

Received : 01 December 2023

Accepted : 31 July 2024

Published : 10 August 2024

DOI : https://doi.org/10.1038/s41467-024-51138-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.