Nanodiamonds were mixed with each protein separately.
Concentration of stock solution was determined by calculate with extinction coefficient of each protein before adding in some amount of nanodiamonds. jNDs and HPHT-NDs are of different size, thus the amount of nanodiamonds used in each sample is different (10mg/mL for jNDs and 1mg/mL for HPHT-NDs) according to their surface size. jND requires high concentration to remain small size and which has to be controlled at this stage. The mixed sample solution was first added in small amount of NaOH or HCl for pH adjustment. Then the sample solution is sonicated for 5 minutes in ice water and shake for another 5 minutes for reaction. Theis process has to be repeated for two more times. After reaction process, the particles were isolated with
ultracentrifuge (CS 150GXL, HITACHI) at 55000rpm for 1.5 hours.
The supernatant was dragged out and the precipitate was washed with 1mL DI water, then shake for 10 minutes before another centrifuge at 55000 rpm for 1.5 hours. The second centrifuge is to obtain the final precipitate and supernatant for complete resolute and further information.
The solution and precipitate at each stage were detected by UV-VIS spectrophotometer(Hitachi U-3310), Delsa™Nano C for particle size and zeta potential.
UV-VIS spectrophotometer were used not only to determine the exceed protein in the supernatant but also the extra nanodiamonds in some conditions which have more nanodiamonds than proteins.
Existence of BSA is determined by the absorption at 279nm, Myoglobin at 409nm and Insulin at 298nm. The amount of proteins are determined by the difference of adsorption peak between supernatant and stock solution. The UV-VIS spectrophotometer of nanodiamond shows a curve line which raises rapidly at lower wavelength, caused by the strong scattering property of nanodiamond.
The condition of ultracentrifuge was set 55000rpm for1.5 hour.
Under this condition both 5nm DSND and 30nm HTHP ND will not have severe precipitate but the mixture of ND and protein. By this way, the mixture particle can be successfully separate from solution.
Delsa™Nano C determined the size of particle by dynamic light scattering(DLS) for controlling and monitoring the condition of particle and supernatant for checking. Zeta potential
Figure F, G: Mixed solution of jND(0-5mg) and BSA(1mg)
Figure H Mixed solution of jND(10mg) and Myoglobin(0.5-5mg)
Figure I Mixed solution of hND(10mg) and BSA(0.025, 0.05, 0.15, 0.2 0.5,1mg, from right to left)
Figure J: Flow chart of react step and centrifuge step
Chapter 4 Results and Discussion
In this work, three types of proteins were chosen to study their static attachment on two different types of nanodiamond: jND with positive surface charge manufactured by Osawa’s group in Japan13 and HTHP-ND (hND) with negative surface charge made in our group.
Additional effort was made to keep NDs from aggregation and, ultimately, achievingthe protein attachment on a nearly single
nanodiamond-size platform. The three chosen proteins were: myoglobin (Mb), insulin (Ins), bovine serine albumin (BSA), each representing a unique pattern of surface attachment on NDs. The pattern of surface adsorption for each protein was characterized by a set of experimental data from UV-VIS spectroscopy, DLS size measurement, zeta-potential measurement and MALDI-TOF, which is discussed in the following sections.
Nanodiamonds without treatments are known to form aggregation, despite the different surface charge they may have carried. Thus, it has been a challenge in keeping the size of ND small and well dispersed when ND is used as a platform to grow protein on its surface. Such ND aggregates of random sizes, which directly affecting protein surface
coverage, present various degrees of difficulties in quantifying the protein-surface interactions. Therefore, small and separated NDs are suitable for this study. Dr. Osawa’s effort in making stable and individual jND with size around 4-6 nm diameter provides an ideal candidate for this work. Some preliminary testing was carried out to determine the stability of jND, in terms of the time duration and pH changes. The experimental conditions used for each protein reported here have all tested in providing individual jNDs of which the size is approaching at the single particle limit. The size of hND, on the other hand, is averaging 30 nm in diameter.
4-1 Myoglobin + ND
4-1-1 Comparison between two kinds of nanodiamonds
The surface interaction of myoglobin with both jND and hND is characterized by the Soret absorption of heme with UV-VIS
spectroscopy. FIGURE 1(A) displays Soret absorption from supernatant of Mb + jND solutions at pH 7.5 in a serial concentration of myoglobin.
The absorption of Mb stock solution at 0.128 mg/mL is included, which is used in deriving Langmuir isotherms. In comparison, another set of
data (not shown here) is also collected from Mb + hND at
FIGURE 1(A): UV-VIS: supernatant of Mb + jND under pH=7.5
FIGURE 1(B): UV-VIS: supernatant of Mb + hND under pH=5
pH 5.0 that shows virtually identical Soret features but at a lower
intensity level. Protocol for spectroscopic measurement is described in the experimental section above, which is similar to previous work of
0
350 400 450 500 550 600 650 700
hND+Mb , supernatant, pH=5
cytochrome c (cyt c) with hND19. The figure shows [jND] fixed at 10 mg for all sample solutions, which is higher than [hND] = 2.0 mg in the counterpart Mb + hND study. The higher jND concentration is
necessary, according to the manufacturer and our own test, to keep jND from aggregation. [Mb] varies from 0.03 to 6.25 mg, corresponding with the number ratio of Mb/jND = 0.05 to 10.5, respectively, covering a range of two orders-of-magnitude change from under surface saturation to well beyond. Some samples in FIGURE 1(A) show measurable absorption signals from Soret band, presumably caused by free myoglobin left in supernatant solution, of which the band intensity
increases as function of [Mb]. Other samples show no measurable Soret absorption, suggesting an under-saturated surface coverage with all Mb attached on jND and no free myoglobin can be found in supernatant. No sign of surface bound myoglobin was observed in supernatant. This assessment is based on the fact that both the Soret peak position (at 409 nm) and band shape remain unchanged in all collected spectra.
Additional evidence comes from the dynamic light scattering (DLS) measurement, discussed later, which shows no surface bound Mb-jND detectable in supernatant, consistent with UV-VIS finding.
Consequently, upon saturating the surface, an equilibrium is established between free Mb in supernatant, Mb(aq), and the surface bound Mb(s) on nanodiamond:
Mb(aq) Mb(s) (Eq. 1)
The surface adsorption constant, Ka, of (Eq. 1), can be determined by the study of adsorption isotherms following the method14. The experimental data fit reasonably well to the isotherm equation of Langmuir’s surface adsorption model (Eq. 2),
Θ = KaCb
1 + KaCb
(Eq. 2) where Θ is the ratio between the occupied surface sites and the total
available sites (when Θ = 1 the surface is saturated) and Cb is protein concentration in each sample solution.
FIGURE 2 shows the derived isotherms for both Mb + jND at pH 7.5 and Mb + hND at pH 5.0. The fitted isotherms are the solid lines in the figure, which give Ka = 8.8×103 and 2.2×105 M-1 for jND and
FIGURE 2: Adsorption Isotherms: Mb + hND & jND
hND, respectively. Therefore, Mb attachment to hND is two orders-of-magnitude thermodynamically more favorable than jND, which is
somewhat unusual. Myoglobin has isoelectric point (pI) at 7 and, at pH 7.5, the protein is very likely to attach on positively charged surface such as jND. On the other hand, at pH 5.0, the negatively charged hND surface is favored. Myoglobin should be attracted to both
nanodiamonds under each set of the experimental conditions in this work.
The difference in the surface adsorption must have come from the size and nature of the two nanodiamonds that eventually makes jND less attractive to Mb. We suspect jND’s small size may have played a
0
Mb Adsorption Isotherms with jND at pH 7.5, hND at pH 5.0
hND + Mb JND + Mb
significant role in this weakening of protein-surface interaction. For one thing, jND (average 4 nm in diameter) compared to hND (average 30 nm) would only have about 1/50 surface area for the incoming protein to contact with. Furthermore, the size of Mb (average 5 nm) is about the same as jND, but much smaller than hND, larger surface areas have proved to enhance the stability in protein-surface interactions. For example, our previous study of cyt c with two different hND sizes, 5 and 100 nm in diameter, shows that Ka-value has increased by 10 times from 5 to 100 nm hND. Another possible factor that may contribute to the large gap in Ka-values is surface charge density, i.e., hND has
substantially higher charge density than jND. When comes to the study of protein surface attachment, it is necessary to consider surface nature (molecular geometry, surface roughness, charge density, etc) and the particle size both ultimately make up the characteristics of each nanodiamond. As will be discussed later, the weak protein-surface interaction of jND becomes ideal in revealing some unique protein-protein interactions upon surface attachment on nanodiamond.
FIGURE 3(a): Zeta-Potentials: Mb + jND
FIGURE 3(b): Zeta-Potentials: Mb + hND
4-1-2 Evidence of the attachment
Now we turn to the protein-nanodiamond complex and some details in their structure. The first question is its existence: Has myoglobin successfully attached to nanodiamond? Electrophoresis experiments were carried out to determine zeta potential of Mb with both nanodiamonds, see both (a) and (b) in FIGURE 3. FIGURE 3(a) shows zeta potential measured from jND (10mg/mL), Mb (10mg/mL) and Mb (0.05mg/mL) + jND (10mg/mL), to be 37.23, -21.24 and 25.73 mV,
37.23
Mb + jND Zeta Potential at pH 7.5
-30
hND + Mb Zeta Potential at pH 5.0
respectively. The positive potential of jND is attributed to its positive surface charge, which has a relatively high value suggesting good stability in the nanodiamond’s individual particle form. Under the
experimental conditions in this work, jND is not likely to form aggregates easily20. Next, the negative potential associated with Mb is expected from its hydrophobicity. The third zeta potential value of 25.73 mV is obtained from mixing small amount of Mb, 0,05 mg, in 1.0 mL of sample solution containing 10 mg jND. This new potential can be viewed as the potential of jND coupling with Mb through static interaction of the two.
Indeed a clear evidence to show presence of the protein-ND complex.
Furthermore, the Mb + jND complex, with a lowered zeta potential, may indicate an increase of the inter-particle interaction between adjacent nanodiamonds, which may lead to aggregation of jND. This is expected to become more significant when protein concentration is raised higher.
FIGURE 3(b) shows a series of Mb + hND samples with increasing protein concentration in 0 – 1.3 mg/mL. Both Mb and hND are negatively charged with respect to the dispersion medium. However, charge coupling the two would raise the potential of hND from nearly -30 mV to virtually zero grounding with respect to the medium at [Mb] = 1.3
mg/mL. At this point, Mb + hND may form clusters in its complex form and hND aggregates, a subject to be discussed later in the DLS study.
FIGURE 4(a) MALDI-TOF: Mb + jND
FIGURE 4(b) MALDI-TOF: Mb + hND
More details of protein surface coverage are provided by the
MALDI-TOF study of protein-nanodiamond attachment. The sample of myoglobin-nanodiamond complex is prepared by centrifuging the sample
0
10000 20000 30000 40000 50000 60000 70000 80000
MALDI: jND + Mb
10000 20000 30000 40000 50000 60000 70000 80000
MALDI: hND + Mb
solution after it has been well mixed then decanting supernatant, washing and centrifuging again. The complex sample goes through MALDI evaporization followed by TOF sorting of fragmentation pattern. See FIGURE 4(a) for the Mb + jND result. One major mass peak appears at 17,120 mass unit, corresponding with the molecular mass of myoglobin, which account for >99% of the total intensity on the mass spectrum.
Trace amount of Mb-dimer can barely be identified at the two times of mass from the parent peak. Similarly, MALDI-TOF spectrum of Mb + hND in FIGURE 4(b) shows only a single parent peak. Two
conclusions can be drawn from this single peak: (1) Mb is indeed
attached to hND, and (2) protein-protein attraction of Mb is weaker than Mb-hND interaction, for the reasons explained below. It is clear that the experimental process described above ensures that all spectral signal obtained must have come from the surface bound Mb. No free Mb can be observed after two runs of centrifuge-and-washing. All mass peaks on the spectrum are indeed resulted from protein on nanodimond surface.
Furthermore, the fact that only the parent peak from Mb on hND observed may indicate Mb-Mb interaction is weaker than Mb-hND attraction. The energy required to detach Mb from the hND surface is
substantially higher than Mb-Mb binding energy and. Therefore, Mb will be fragmented to form single particles in the process of protein detachment from the surface and only one peak will show on the mass spectrum. By the same argument, the surface attraction of jND
apparently is not as strong as hND and, through only a trace amount, Mb-dimer can still survive after detached from jND.
4-2 Insulin + ND
4-2-1 Comparison between two kinds of nanodiamonds
Insulin (Ins) has isoelectric point at 5.4, therefore, all Ins + jND experiments were carried out at pH 6.0 in this work and Ins + hND at pH 4.5. FIGURE 5(a) shows UV-VIS absorption of Ins + jND supernatant from sample solutions in a series of insulin concentration with [jND] all fixed at 10 mg/mL. A 1.0 mg/mL insulin stock solution is also included to show the absorption band and peak position around 277 nm. The samples in the figure contain a concentration range
FIGURE 5(a) UV-VIS Ins + jND
of [Ins] = 0.5 - 10 mg/mL corresponding with the (Ins/jND) number ratios of 2.5 - 50, respectively. At lower [Ins], free nanodiamond particles appear present in supernatant, causing scattered light that is more sensitive towards the far UV region. Therefore, as seen in the figure, the signal intensity of [Ins] = 0.5 and 2.0 mg/mL is tilted higher in the lower wavelength region. As more insulin added, the signal flattens out eventually becoming indistinguishable from baseline. Interestingly, no sign of extra free insulin is visible in the figure even at the highest [Ins] (10 mg/mL) where (Ins/jND) number ratio reaches at 50. This is quite different from the Mb + jND study reported above where free Mb particles are clearly present in supernatant when the jND surface is saturated. Is it because Ins has not saturated the jND surface? Or,
-0.1 0.4 0.9
270 280 290 300 310 320 330 340 350
Supernatant of jND+INS
could it be a sign of a different surface attachment? Not a simple thin layer formation on the surface, rather, some other form of protein
assembly linked to the surface? To explore this issue, we compare the hND surface attachment by insulin.
FIGURE 5(b), upper graph, Ins+hND at low [Ins], 5(c), lower graph, at high [Ins]
-0.1
270 280 290 300 310 320 330 340 350
hND+INS, pH=4.5, Supernatant
270 280 290 300 310 320 330 340 350
hND+INS_pH=4.5 , supernatant
FIGURE 5(b) shows insulin absorption of some Ins + hND
supernatant samples consisting of low [Ins] (upper graph) and high [Ins]
(lower graph). With [hND] fixed at 1.0 mg/mL, the samples in the upper graph contain [Ins] ranging from 0.01-0.05 mg/mL corresponding with the (Ins/hND) number ratios of 1-5, respectively. Similar to the surface attachment on jND, FIGURE 5(a), the spectrum features only free hND scattered light visible in the far UV region, which decreases in increasing Insulin concentration. At [Ins] = 0.05 mg/mL, the amount of scattered light becomes small enough that the signal intensity virtually merges with baseline of the spectrum. Presumably, at this
concentration, most of hND particles are precipitated out by insulin in some form of Ins-hND complex leaving no free hND in supernatant.
Continuing in raising Insulin concentration, the lower graph in FIGURE 5(b) shows free insulin beginning present in supernatant.
While no sign of insulin is observed at the least [Ins] (0.1 mg/mL) in this group of samples, a distinctive absorption band emerges at around 280 nm as [Ins] grows to 0.5 mg/mL and more. The band matches
identically with both the shape and peak position to the absorption of the insulin stock shown in FIGURE 5(a), which verifies the presence of free
insulin in supernatant. Furthermore, the absorption band intensity
increases when Insulin concentration raises in supernatant, the same as in the case of Mb + hND. It seems reasonable to speculate that the pattern of surface attachment on hND is similar for the two proteins. Like myoglobin, beyond surface saturation of hND, equilibrium is established between insulin particles attached to the surface and free in solution.
The surface saturation point, which may be experimentally determined as at the sample concentration where UV-VIS spectrum shows no free hND or free insulin present in supernatant. At this point, the signal merges with baseline in the spectrum. For Ins + hND, the saturation points occurs at [Ins] = 0.1 mg/mL in FIGURE 5(b). The associated number ratio is 10, presumably, each 30 nm-dia. hND can accommodate10 insulin particles on its surface. Previous work19 found an hND particle of 5 dia. can carry 2 cytochrome c (average 3 nm-dia.). Considering the size differences of both nanodiamond and protein, average 4 nm for insulin, it is quite feasible that 30 nm-dia.
hND would be able to carry 10 insulin particles. Of course, the surface-protein interaction must be about the same for these two surface-proteins.
While hND seems to show similar surface interaction with both Mb and Ins, based on the study of their UV-VIS data above, it is not clear if jND would have the same effect on both proteins. In fact, there is significant discrepancy between how Mb and Ins attaching to jND.
FIGURE 1(A) shows free Mb present in supernatant when reaching jND surface saturation and beyond. This is never observed in Ins + jND, even with the (Ins/jND) number ration at as high as 50 in FIGURE 5(a).
So, how is insulin attached to jND? To resolve this mystery, we look to mass spectroscopy for clues.
The capability of nanodiamond to capture protein from solution was demonstrated by our previous work with three different proteins:
bovine cytochrome c, myoglobin and bovine serine albumin14. Using MALDI-TOF, proteins were shown to be detectable in the quantity as small as femto mol, limited mostly by the instrumental sensitivity.
Following this method, insulin on both jND and hND was measured and the results were included in FIGURE 6(a)-(c).
FIGURE 6(a) MALDI-TOF: Ins + hND
FIGURE 6(b) MALDI-TOF: Ins + jND, low [Ins]
0
5000 10000 15000 20000 25000 30000 35000 40000
0
FIGURE 6(c) MALDI-TOF: Ins + jND, high [Ins]
FIGURE 6(a) shows insulin mass peaks obtained from Ins + hND prepared in 1.0 mg/mL of hND and 2.0 mg/mL of Ins at pH 4.5. The mass spectrum consists of a dominating peak at around 5800,
corresponding with insulin’s molecular mass, and its dimer peak with intensity less than 1% of the monomer. At this concentration, the
corresponding Ins/hND number ratio is 100, which is well above surface saturation, 10, as estimated above. Additional samples at higher Insulin concentrations, up to 10 mg/mL, all show the same pattern of mass peaks.
Therefore, similar to Mb, the mass data of Ins + hND seem to support the model of a monolayer surface coverage on hND. At surface saturation,
0
5000 15000 25000 35000 45000 55000 65000
MALDI-TOF: Ins + jND (10 mg, 10 mg)
equilibrium is established between insulin attached to the surface, Ins(s), and free in solution, Ins(aq):
Ins(aq) Ins(s) (Eq. 4)
on hND. This surface attachment pattern is a result of strong surface-protein interaction, which reduces the surface-protein-surface-protein interaction, preventing proteins from stacking on more than monolayer.
Insulin attachment on jND has given mass spectra in FIGURE 6 (b) and (c). The two samples represent two different levels of insulin
content: the lower [Ins] (4 mg/mL) in (b) and the higher (10 mg/mL) in (c), respectively. Unlike hND, there are multiple mass peaks occurring in the spectrum, each corresponding with mass of a multimer of insulin.
In the zoom-in of FIGURE 6(b), the sample contains [Ins] = 4 mg/mL showing mass peaks up to 4 times of the insulin molecular mass.
In the zoom-in of FIGURE 6(b), the sample contains [Ins] = 4 mg/mL showing mass peaks up to 4 times of the insulin molecular mass.