Functionalized graphene-based biosensors for early detection of subclinical ketosis in dairy cows

Precision livestock farming utilizing advanced diagnostic tools including biosensors can play a key role in the management of livestock operations to improve the productivity, health, and well-being of animals. . Detection of ketosis, a metabolic disease that occurs in early lactation dairy cows due to the negative energy balance, is one potential on-farm use of biosensors. Beta–hydroxybutyrate (βHB) is an excellent biomarker for monitoring ketosis in dairy cows because βHB is one of the main ketones produced during this metabolic state. In this report, we develop a low-cost, Keto-sensor (graphene-based sensor) for the detection and quantification of βHB concentrations in less than a minute. In this device, graphene nanosheets were layered onto a screen–printed electrode (SPE), and then a stabilized enzyme (Beta–hydroxybutyrate dehydrogenase, NADH, and glycerol) was used to functionalize the graphene surface enabled by EDC– NHS conjugation chemistry. The Keto-sensor offers an analytical sensitivity of 10 n M and a limit-of-detection (LoD) of 0.24 n M within a detection range of 0.00001-3.0 m M . Spike testing indicates that the Keto-sensor can detect βHB in serum samples from bovines with subclinical ketosis. The Keto-sensor developed in this study shows promising results for early detection of subclinical ketosis on farms.


Introduction
The tools of precision livestock farming offer opportunities to monitor the health status of dairy herds with early detection of disease 1 .On-site sensing technologies may identify early signs of illness minimizing cost and time.They may provide actionable decision-making tools for the management of dairy herds 2 .
Subclinical ketosis (SCK) 3,4 is an important metabolic disease in early lactation cows caused by intense demand for glucose and the mobilization of adipose tissue during a cow's transition period 4 .While there is no visual indication of ketosis, elevated ketone concentrations (acetone (Ac), acetoacetate (AcAc), and beta-hydroxybutyrate (βHB)) in bodily fluids such as milk and blood are markers of SCK 5 .If not treated, a ketotic cow will lose its appetite, have decreased production and reproductive performance, and become more susceptible to other diseases such as mastitis, displaced abomasum, and metritis 6 .It is estimated that the prevalence of SCK is high (~40-60%) compared to clinical ketosis (~2-15%) 7 and prevention is valuable because the cost of SCK includes the treatment of the animal, loss of milk production, and delays in conception.Estimates of costs per case vary, ranging from $24 to $1030 per case 8 , and averaging $165 per case based on stochastic simulation modeling 9 .Rapid detection of SCK at its earliest stages allows the producer to make better management decisions for their animals and minimize the disease's economic impact.
The quantification of βHB is known as a gold standard for the diagnosis of SCK in dairy cows 10 .A dairy cow with concentrations of blood βHB greater than 1.0 mM is considered to have SCK 11 .Standard thresholds for SCK are 100 µM for milk and 1.5 mM for urine 11 .Traditional diagnostic tools for ketosis include cow-side urine dipstick tests and laboratory-based diagnostic tests.Dipstick tests involve the stimulation of urination on a stick of paper 4,12 .The color change on the paper reflects the quantity of ketones present in the cow's urine.This cow-side test is simple but neither precise nor accurate.Another common tool to detect ketosis involves sending samples of blood, urine, or milk from cows to a laboratory to be analyzed using enzymatic tests based on spectrophotometry 11,13 .Such assays are accurate 14 but sample processing and shipping from herds to testing facilities limit their utility for use in making management and treatment decisions.To address these challenges, on-site biosensing tools could be helpful in the management of cow health for their rapid and early diagnosis of SCK 10 .Accurate, sensitive, and precise biosensing tools would enable quick decisions, on-farm detection, prevention of clinical ketosis, reduce economic losses of the disease, and improve the well-being of cattle.A microfluidic biosensor developed by researchers was able to detect SCK in one minute with sensitivities in the range of millimolar concentrations (i.e.0.05 mM) 10,15 .These biosensor tests use quantum dots 15 and microfluidics with optical coupling 10 .Additional ketosis tests developed by Ketolac (Biolab, München, Germany) 16 , and Optimum Xceed (Voyvoda and Erdogan) 17 are commercially available.Such hand-handled devices were made https://doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0 exclusively for human monitoring; thus, they have limited sensitivities (less than 85%) compared to regular laboratory tests.One other commercial tool was developed by Novavet 18 for the detection of ketosis; it has limited sensitivity of 82% along with significant false-positive results.Thus, there is an unmet need to develop reliable and low-cost ketosis biosensors that can allow for the on-farm detection of SCK in dairy cows.
Nanostructures with two-dimensional geometries are exploited to construct advanced biosensors to track many biomarkers in biological media 19 .This allows surface functionalization to hold enzymes or antibodies due to their high surface area 20 , and also enhances the electron transport properties, resulting in rapid detection of the target compound.2D materials, specifically graphene nanosheets, provide abundant functional groups (-COOH) to bind with enzyme molecules (-NH2) covalently via amidation reactions 21,22 .However, the enzymes need to have high stability on the sensor surface for on-farm detection 23 .
Graphene-based biosensors have been used in agriculture to detect microorganisms in food, monitor crop health, and measure biomarkers of disease in livestock but unstable enzymes limit performance 24 .Thus, enzyme stabilization is a key to success for commercial field-testing and wearable sensors 25,26 providing advantages such as a) increased sensor-to-sensor reproducibility and sensor shelf-life, b) reduced biofouling, c) maintenance of sensor functionalities, and d) reduced tendency of the enzyme to unfold 25,27 .
In this study, we have developed a low-cost, and highly sensitive nano-enabled electrochemical biosensor (i.e.Keto-sensor) that detects βHB to identify SCK in dairy cows.This Keto-sensor consists of a screen-printed electrode (SPE) which is modified with graphene nanosheets to allow covalent functionalization aided by N-Ethyl-N′-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxy succinimide (NHS) chemistry of enzymes.The sensor utilizes two enzymes, beta-hydroxybutyrate dehydrogenase (βHBD) and β-nicotinamide adenine (NADH), which were stabilized with a glycerol treatment before their immobilization.These enzymes on the sensor surface allow electrocatalytic reactions, and generate electrons to be measured via a graphene integrated SPE current collector.As expected, enzyme stabilization improved sensor stability and selectivity.As a control, the Keto-sensor was tested against a screen-printed sensor (S-sensor) without graphene or enzymes and a sensor with graphene nanosheets (Gsensor) without enzymes.We conducted a systematic sensing evaluation of all sensors (S-sensor, G-sensor and Keto-sensor).Upon calibrations, the spiking analysis with complex biological matrices including bovine serum, the biosensor showed a good validation.This sensor can detect βHB within a minute and is sensitive to a nanomolar (0.24 nM) concentration of βHB.This sensor also demonstrates the ability of continuous monitoring of βHB concentration.Results show a promising alternative diagnostic tool that detects SCK in dairy cows.

Chemicals and reagents
Graphene oxide (ACS Material, CA, USA) was used to layer for enzyme functionalization onto the WE of the G-sensor and Keto-sensor.Phosphate buffer saline (PBS) solution was used to prepare stock solutions of NHS, EDC, βHBD, NADH, and βHB, as well as the buffer solution to establish a sensor baseline.Fetal bovine serum (FBS, Sigma Aldrich, MO, USA) was used for spiked sample testing as serum could be a potential matrix to measure βHB.βHB was used to make standard titration solutions (0.01-3000 µM) to calibrate the Keto-sensor.βHBD, a specific enzyme to βHB, was functionalized onto the graphene surface via EDC-NHS chemistry.Within the enzyme solution were NADH and glycerol.NADH takes part in the oxidation-reduction reaction that occurs when βHBD catalyzes βHB and glycerol is a stabilizing agent for βHBD.All chemicals such as βHBD, βHB, NADH, EDC, NHS, and glycerol were purchased from Sigma Aldrich, MO, USA.Further explanation of how stock solutions were prepared can be found in the Supporting Information (see Section 1).

Instrumentation
SPEs (BASi, Inc., IN. USA) were used to build sensors.The use of commercial electrodes avoided cleanroom fabrication, reduced the device cost, and ensured the reproducibility of the sensor.Further, these sensors can easily be interfaced with a potentiostat readout for collections via Bluetooth.Sensors were inserted into a commercial readout (EmStat Blue, Palm Sens, Netherlands) to conduct experiments.The Bluetooth-enabled potentiostat connected via Bluetooth to a computer or tablet to collect data (PSTrace, Palm Sens, Netherlands) and data were exported to Origin.Inc (OriginLab, MA) to create graphs.
BioRender was used to create the schematics in Figure 1.
Electrodes are screen-printed onto a paper substrate.The WE for this keto-sensor was modified.A photo of a keto-sensor is shown in Fig. 1A.A solution containing highly dispersed 2D graphene nanosheets was pipetted onto the WE and dried for one hour at 80°C.This procedure was done twice for a uniform surface of graphene.In this process, a thin layer of graphene nanosheets was layered due the non-covalent π-π interactions of graphene and carbon 28 .Next, EDC and NHS solutions were applied at a one-to-one ratio to the WE and the sensor was placed in a humid chamber for four hours 29 .After four hours, the electrode was washed with commercially available PBS (Gibco, MA), and the sensor was functionalized with enzymes.This EDC-NHS treatment 30 activated the abundant -COOH groups on the graphene modified Keto-sensor to bind with proteins.
https://doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0 For enzyme functionalization, an enzyme solution was separately prepared consisting of NADH and βHBD mixed at a 1:1 ratio.To stabilize the enzyme, glycerol was mixed into the final enzyme solution as 5% of the final volume.20 µL of this solution was spread uniformly on the surface of the graphene electrode.In this EDC-NHS chemistry, -COOH groups on the graphene surface bind to -NH2 of the enzyme and form a C-N covalent bond via an amidation reaction 21 .In brief, the EDC reacted with the graphene -COOH groups and formed o-acylisourea which can immediately react with NHS of the enzyme molecules resulting in NHS esters.These NHS esters can react with amines of the enzyme to form a C-N covalent bond 21 .These functionalization steps are visualized in Fig. 1D.The sensor was placed in a humid chamber for at least four hours but no longer than twelve hours.After incubation, the electrode was washed again with PBS, then the sensor was placed in a 4ºC refrigerator until use.
The potentiostat connects the three connectors that lead to the CE, WE, and RE as seen in Figure 1A.On the WE are layers of graphene with enzyme, represented by the scanning electron microscopy (SEM) images in Figure 1B.A cow's blood glucose concentrations decrease immediately after calving because of the negative energy balance caused by the sudden onset of lactation and the limited food consumption typically observed.This negative energy balance triggers catabolism of body stores of fat to provides energy, and ketone bodies like βHB are produced as a byproduct.Elevated blood ketone concentrations further impair feed intake, and this cycle often leads to SCK.This process, of a dairy cow developing ketosis after calving, is outlined in Figure 1C 31 .Graphene oxide was added to the screenprinted WE, EDC-NHS coupling with the graphene was done as described above, the enzyme solution was added to the sensor, and finally, sensing was performed with this functionalized Keto-sensor.βHBD catalyzes the βHB to acetoacetate.The role of NADH in this reaction is to act as a reductant for the reaction creating βHB and NAD+ acts as an oxidant when the reaction moves from βHB to acetoacetate 32 .The glycerol in the enzyme solution ensures stability over time 33 .Milk, blood, and urine samples can be collected from the cows and used to monitor βHB as outlined in Figure 1E.In addition to the Keto-sensor, two more sensors were chosen as controls in this study.These control sensors were the S-sensor (SPEbased sensor) and the G-sensor (graphene modified screen-printed sensor); these sensors do not contain specific enzymes i.e. βHBD.
https://doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0 Once the graphene was dried onto the surface of the substrate, then an EDC-NHS solution was added and allowed to sit in a humid chamber.After 4 hours, the excess EDC-NHS was washed from the surface and the βHBD enzyme solution was added and allowed to sit in the humid chamber for up to 12 hours.When testing the sensor, the βHB in the sample reacted with the βHBD.(E) Visual representation of a dairy cow producing milk that can be tested for βHB as a biomarker of SCK.

Surface morphologies
To investigate the surface morphologies of the sensors, we conducted scanning electron microscopy (SEM) imaging along with energy-dispersive X-ray spectroscopy (EDS) using the JEOL IT500 SEM and Oxford Instrument AZtechOne Detector.The SEM imaging was conducted for the screen-printed sensor without modification aka S-sensor (Fig. 2A), the graphene sensor without enzyme aka G-sensor (Fig. 2B), and the graphene nanosheets along with enzyme aka Keto-sensor, (Fig. 2C) at 500X.The same sensors used as SEM samples were coated with a thin layer of iridium for EDS analysis.The SPE showed the bare bulky carbon structure that is packed together with a non-uniform surface (Fig. 2A).This morphology was changed with nanosheets of graphene layers (Fig. 2B).These nanosheets were seen to be connected and formed a porous, thick layer.This porous layer was expected due to the π-π interactions among the nanosheets of carbon.Some nanosheets formed wrinkles due to their stacking and folding.This nanoenabled sensor surface not only increased the area of surface reactions but also enhanced the loading of enzymes.The morphology was further changed when enzymes were immobilized via EDC-NHS chemistry (Fig. 2C).The enzyme layer on graphene, however, was unclear to observe., 2E, and 2F display the mapping results from the EDS of the Keto-sensor to evaluate individual elements.The EDS spectrum of the Keto-sensor indicates the presence of respective elements (Fig. 2G).The surface was completely covered in carbon (C~90.9%)as the SPE uses carbon (Fig. 2D).

Figures 2D
Oxygen (O) was dispersed across the surface of the SPE as the graphene nanosheets applied to the WE were made up of graphene oxide (Fig. 2D).The distribution of nitrogen (N) on the WE correlate to where the enzyme was attached to the graphene layer (Fig. 2F).The presence of N indicated enzyme immobilization on the sensor surface.The addition of graphene nanosheets to both the Keto-sensor and Gsensor was likely the cause of the increase in the O weight percentage (wt%) of the WE as compared to the S-sensor seen in Figure 2H.N was present on its WE, but not on the WEs of the S-sensor and G-sensor due to the addition of βHBD to the Keto-sensor (Fig. 2H).A table comparing the three different sensors is shown in Fig. 2H.

Electrochemical sensing of ketosis and characterization
Electrochemical studies were conducted to investigate the redox properties of the sensors: the Keto-sensor, S-sensor, and G-sensor.The cyclic voltammetry (CV) measurements of all sensors were conducted using a PBS solution containing an equimolar concentration (5 mM) of a ferro/ferricyanide redox mediator.The cyclic voltammograms of the Keto-sensor in the absence and presence of 1mM βHB as the target analyte demonstrate the ability of the Keto-sensor to detect βHB (Fig. 3A).A pair of oxidation and reduction peaks in both graphs was related to the redox reaction of the mediator on both sensors.However, in the presence of βHB (1 mM), the Keto-sensor showed a decrease in the redox peaks and a significant shift towards greater oxidative potential.Additionally, a dominant oxidation peak appeared at 0.37 V (vs.Ag/AgCl) in the presence of βHB.This peak is considered the sensing signal for the measurements of βHB concentration.
As the Keto-sensor contains enzymes that catalyze the βHB to acetoacetate and produce electrons at an oxidation potential of 0.37 V, this confirms that the Keto-sensor selectively detected the presence of βHB https://doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0 in buffer solutions.Another peak was observed at nearly 0 V due to the unreacted EDC-NHS on the surface of the graphene that was used in the sensor construction.The CV tests showed a clear difference between the S-sensor, G-sensor, and Keto-sensor in the presence of 1 mM βHB (Fig. 3B).The S-sensor was a bare SPE sensor without any modification.This sensor showed strong oxidation and reduction peaks in the presence of βHB (1 mM) in the buffer solution which was attributed to the redox peaks of ferro/ferricyanide mediators.The oxidative peak potential and current were at ~0.42V and 150 µA.When the electrode surface was modified with graphene nanosheets (G-sensor), this mediator oxidation peak potential decreased to 0.3V and the current decreased (50 µA) significantly.This lower current was due to the available functional groups at the graphene surface that block the electron transfer from the mediator to the electrode.The differing performance of the G-sensor and S-sensor in detection of βHB concentration is presented in Figure 4. Though the current was lowered with the G-sensor, this configuration was used to add the stabilized enzyme to the WE due to its lower sensing potential and ability to enhance the selectivity of the sensor.In more oxidative potentials, the electro-oxidation of interfering molecules in the complex matrix of the real sample can interfere with the actual detection of βHB.The Keto-sensor had notable multiple oxidation and reduction peaks in comparison with the S-sensor and G-sensor (Fig. 3B).The addition of the enzyme on the Keto-sensor created these additional peaks, indicating that the enzyme was reacting with the analyte causing the exchange of electrons, resulting in high selectivity.Ultimately, the Keto-sensor boosts the sensing signal more than three times, even if there was an enzyme coating on the sensor, as compared to the G-sensor that can diminish the redox reactions due to the insulative effects of enzymes.The promoted electron exchange properties of βHBD were due to the enzymatic reactions during the detection of βHB concentration.Unlike the Keto-sensor, the G-sensor and S-sensor did not show peaks other than the mediator redox peaks in their respective CV graphs (Fig. 3B during sensing of βHB.
A scan rate study was performed using the Keto-sensor.For this study, a range of potential from -0.7 V to 0.7 V (vs.Ag/AgCl) was applied at different potential scan rates.The results of scan rate studies that were conducted using the Keto-sensor and tested scan rates of 0.5 V/s, 0.75 V/s, and 1.0 V/s are displayed in Figures 3C and 3D.With the increase in scan rate, the oxidation peaks during the cyclic voltammetry increased, indicating the surface-controlled process of the Keto-sensor.With a change in scan rate, there was a shift in the oxidation peak current to the right and a shift in the reduction peak current to the left with increasing scan rate that increases the ΔEp.The peak currents were linearly proportional to the scan rate.
The sensing performance of all three sensors was investigated using electrochemical characterizations.
Though the Keto-sensor provided comparable sensing with the S-sensor, it is expected that the Keto-sensor can provide more selectivity due to the added enzyme layer that can generate electrons directly from the electro-oxidation of βHB species.The lower sensing potential of the Keto-sensor is another advantage compared to the S-sensor that can boost selectivity.

Ketosis (βHB) sensing
Enzymatic and non-enzymatic sensing performances were investigated for the S-, G-and Keto-sensor.In the non-enzymatic sensing modality 32 , the S-and G-sensors performed electrocatalytic reactions without any enzymes, while the enzymatic Keto-sensor 34 performed enzymatic oxidation resulting in electron generation in the presence of βHB.Carbon 35 and graphene 36 can both act as electrocatalysts.These are useful materials to construct non-enzymatic sensors but the selectivity in their performance is limited in comparison with enzymatic sensors.This causes challenges in complex sample matrices like milk or serum.
Error bars were calculated by calculating the standard deviation of three measurements.
βHB sensing results for S-sensor were collected using dose-dependent concentrations (Fig. 4A-B) using CV measurements (Fig. 4A).Initially, the S-sensor was exposed to a buffer solution to set the sensor baseline.Clear oxidation and reduction peaks were observed due to the presence of a ferro/ferricyanide mediator.The oxidation current was found to be at 155 µA.Next, a minimum concentration of βHB (0.01µM) was introduced to the S-sensor.The peak current was enhanced significantly compared to the sensor's baseline.This is due to the non-enzymatic electrocatalysis oxidation of βHB molecules on the carbon surface.Then, the S-sensor was washed with buffer solution before the next solution with a higher concentration was introduced.The βHB concentration was increased from 0.1µM to 3 mM and signal increases (peak current) were observed directly proportional to the increased concentrations of βHB (Fig. 4B).However, the S-sensor signal became saturated after 0.5 mM concentration of βHB.This is one of the major limitations of the S-sensor and is overcome by introducing graphene layers along with a stabilized enzyme on the sensor surface.From the sensor calibration, the slope value was estimated as ~ 48.8 ± 1.3 µA.
Similarly, sensing measurements were conducted for the G-sensor in the presence of all standard target concentrations of βHB (Fig. 4C-D).On adding a graphene layer, the oxidation peak potential was drastically reduced to 0.2 V. Furthermore, the peak current of the baseline's G-sensor was also reduced to 33 µA.The lesser baseline signal was due to the presence of functional groups at graphene sheets that impede the electro-oxidation of βHB on its surface 38 .Then, a low concentration of βHB (0.01µM) was introduced, and the peak current of the G-sensor was increased notably.urther, the concentration of βHB was increased and the peak current was increased at a potential of 0.2 V.The peak current was found to increase until a 0.5 mM concentration of βHB, after which the -sensor showed saturated results similar to those observed with the S-sensor.For this sensor, the slope of the calibration was estimated as ~ 26.6 ± 2.6 µA (Fig. 4D).The limit of detections (LoDs) for both the S-sensor and G-sensor were estimated as 857.02 nM, and 545.56 nM, respectively.The detailed calculation of LoD is demonstrated in the Supporting Information (see Section SI).The analytical sensitivities of the S-and G-sensors were 0.01 µM, but these sensors were not sensitive beyond 0.5 mM concentration of βHB.
https://doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0 monitor animal health with clear differentiation between healthy cows and those with SCK.This monitoring would allow management decisions to both prevent and treat the disease.Table I compares the Keto-sensor created during this study with sensors developed by others for detecting βHB.Many of the devices made to measure βHB focus on use with human patients, however, this device and one other 41 were made with the intended goal of on-farm use.Many of these devices used amperometry for measuring [41][42][43] , while this device and one other 44 used cyclic voltammetry measurements.
The limit of detection for the Keto-sensor was 0.24 nanomolar, while the limit of detection for the G-sensor and S-sensor were 545.56 nanomolar and 857.02 nanomolar, respectively.The Keto-sensor has a much lower limit of detection than all other sensors (Table I).The wide range of detection (0.00001 to 0.

Selectivity studies
For the selectivity test, potential co-existing target molecules available in serum samples of dairy cows were applied to the sensors (Fig. 5D-E).These molecules can undergo a non-enzymatic electro-oxidation reaction and interfere with the βHB detection.Glucose (2 mg/mL; Sigma-Aldrich, MO) and urea (2.5 mg/mL; Fisher Chemical, MA) solutions, with no or 1 mM added βHB were used to test for selectivity.
First, the Keto-sensor was set to the baseline.On exposure to glucose and urea, the Keto-sensor did not show a significant change, as expected.The Keto-sensor provided a sensing signal when βHB was added.
hen βHB was added to glucose and urea solutions, the Keto-sensor showed a slight change in response current.However, the relative standard deviation (RSD) was estimated as ±9.1%.Such low RSD indicated that the Keto-sensor was selective, able to detect βHB even in the presence of other compounds.

Conclusions
In summary, this Keto-sensor shows promising results as it can detect both clinical and SCK in the serum of dairy cows with a response time of less than a minute.Detecting βHB at such a low concentration (0.01 µM) will allow farmers to monitor the changes in their cows' metabolism before any problems arise.Additionally, the fast response time is ideal for field use of this sensor.This Keto-sensor also shows promise in the lab setting displaying differences between samples spiked and not spiked without βHB.Additionally, testing continuous measurement over longer periods will allow continuous ketone sensing.Selectivity testing should continue with more biological compounds to further prove the good selectivity of the ketosis sensor.The limitations to on-farm use of this device include the need for proper training of farmers in its use and interpretation of results.Precision agriculture management is the future as the world needs to produce more food using less land and fewer animals for our growing population 45 , and precise, accurate biosensors like the one developed will help support sustainable agricultural production across the globe.

Figure 1 .
Figure 1.Schematic of a Keto-sensor to measure concentration in dairy cows.(A) Keto-sensor setup displaying the connectors, working electrode (WE), counter electrode (CE), and reference electrode (RE).(B) SEM image of the enzyme that was immobilized onto the graphene/WE.(C) Visual representation of how decreased glucose levels lead to elevated fat metabolism and increased circulating βHB concentrations in dairy cattle.(D) Schematic showing the creation of the WE for the Keto-sensor.Functionalized graphene coats the carbon substrate of the screen-printed electrode (SPE).Once the graphene was dried onto the surface of the substrate, then an EDC-NHS solution was added and allowed to sit in a humid chamber.After 4 hours, the excess EDC-NHS was washed from the surface and the /doi.org/10.26434/chemrxiv-2024-zj1j2ORCID: https://orcid.org/0000-0001-5752-8808Content not peer-reviewed by ChemRxiv.License: CC BY-NC 4.0

Figure 2 .
Figure 2. Investigations of the surface morphologies of the sensors.SEM images for screen-printed sensor without modification (A), graphene sensor without enzyme (B) and graphene nanosheets along with enzyme or Keto-sensor (C).(D-F) EDS mapping of the WE of the Keto-sensor showing carbon, oxygen, and nitrogen distribution.(G) Graph showing weight distribution of different elements found on the WE of the Keto-sensor.(H) Table showing the weight distribution of the different elements found on the WE of the ketosis sensor.

Figure 3 .
Figure 3. Electrochemical characterization of the sensors.(A) Cyclic voltammetry (CV) of the Ketosensor with and without βHB (1 mM) concentration.For the CV test, the PBS solution contains a 5 mM concentration of a [Fe(CN)6] 3−/ − redox mediator.(B) CV comparison for the S-sensor, G-sensor, and Ketosensor in the presence of βHB (1 mM) concentration.(C) Scan rate study for Keto-sensor wherein the scan rate was varied from 0.5 V/s to 1.0 V/s.As the scan rate increases, the peak currents increase.(D) The calibration plot of scan rate studies compares the peak current during different scan rates.

Figure 4 .
Figure 4. Sensing performance of the sensor.All dilution studies were conducted at room temperature and used the same βHB serial dilutions, and buffer solution (0.01 µM, 1.00 µM, 0.10 mM, 0.25 mM, 0.50 mM, 1.00 mM, and 3.00 mM) for all the sensors.(A-B), (C-D) and (E-F) present the sensing performance of the S-sensor, G-sensor, and Keto-sensor, respectively.A, C and E show the CV responses for the Ssensor, G-sensor, and Keto-sensor, respectively, with increasing concentrations of βHB in buffer solution.B, D, and F are the sensor calibration plots for S-sensor, G-sensor and Keto-sensor, respectively.In all three

Figure 5 .
Figure 5.Studies with real samples, continuous, and selectivity measurements.(A) A comparison plot showing the calibration plots for S-sensor, G-sensor and Keto-sensor.(B) Cyclic Voltammetry graph of Keto-sensor testing deionized water, fetal bovine serum, and fetal bovine serum spiked with 3 mM of βHB.The addition of βHB to the serum increases the peak values recorded during CV.(C) Chronoamperometry graph showing long-term and continuous sensing of 3 mM βHB, 1 mM βHB, and buffer solutions.For this study, we continuously ran 10 min of measurements.(D) graph for selectivity study on the Keto-sensor.Different solutions of urea, and glucose excluding and including βHB were added to the sensor and chronoamperometry was performed for one minute (60 seconds) with each solution.Between Figure 5D demonstrates the chronoamperometric graphs and Figure 5E is a bar graph made from Figure 5D to show the difference in current measured from each solution.The stabilized enzyme was responsible for such selectivity as this enzyme can only have biochemical reactions with βHB during detection.

Table I .
1 mM and 0.25 to 3.0 mM) allows the Keto-sensor to differentiate between healthy dairy cows and those with SCK and clinical ketosis.This capability shows the advantages of the Keto-sensor using graphene nanosheets with a stabilized enzyme on the WE.Further, long-term and continuous measurement of βHB is another important feature of the Keto-sensor compared to others reported in the literature.Comparison table shows the sensor performance with other report literature.