quantitative fluorescence of nuclear DNA is still a mystery

........................................................................................................................................

The cytometrical usage of DAPI

The following issues will be discussed:

1.  Binding of DAPI to DNA.
2.  The Jablonski diagram.
3.  Long time fluorescence fading elucidates chromatin fractions differing in the degree of
     condensation
4.  Short time fluorescence fading exhibits information on binding mechanisms.
5.  How many photons will excite one DAPI molecule in a low power flow cytometer?
6.  Few DAPI molecules are surrounded by a predominance of binding sites.
7. A least comment will close this session.

Some conclusions are given beforehand:

1. AT repeats:  DAPI binds to the minor groove with pronounced enhancement of fluorescence.
    GC repeats:  DAPI binds intercalatively with narrow fluorescence such as the free dye.
3. Long time photo-bleaching over minutes exhibits a superposition of three exponentials and allows
    the quantitation of eu- and heterochromatin.
4. Short time fluorescence fading demonstrates three binding modes with a main decay time of 4
    nanoseconds.
5. Excitation is poor in a low power (10 mWatt) flow cytometer and is confined to one photon per
    each DAPI .
6. Depending on the level of chromatin configuration the ratio of binding sites to DAPI ranges from
   70 to 320 with a respective cubic spacing of binding sites from 13 to 8 nanometer.
7. Assuming a nucleus of 6 billion base pairs DNA (6pg and a GC/AT ratio of 1), roughly 6 million
    binding sites will be accessible for DAPI to bind, which is 1000 to 1.
                                           ..................................................


  • 1.  Binding of DAPI to DNA

 
Quantitative cytochemistry of cellular DNA combined with cytometric technology is a sophisticated
type of cytochemical balance for estimating the amount (mass in picograms) of DNA in cells.

It will be shown that the dyes utilized for this job are bound only to a very small portion (1% or lower) of total DNA, and the fundamental concept of quantitative staining makes the assumption that this small part relates to total DNA in a constant proportion. This is like drawing conclusions from the weight of the hair coat to the whole body mass.
Most protocols for fluorescent staining of cellular DNA have been developed in the 70’s and 80’s of the last century.

Nowadays a lot of kits are available for this purpose and the actual laboratory knowledge about the
binding mechanisms increasingly fades away and ignorance spreads out to some degree.
Nevertheless, quantitative DNA cytochemistry for providing a cell-by-cell estimate of the amount and organisation of nuclear chromatin, is still a mystery. The following considerations are trying to elucidate some aspects of dye binding to double stranded DNA in cell nuclei.

DAPI (Dann et al. 1971) was chosen as a representive molecule for this approach, but the results are
in some respect transferable to other stains.

The molecular structure of DAPI is shown with Fig.1.














Fig. 1:
Molecular structure of DAPI

DAPI is an analogue of diarylamidine (berenyl). It exhibits no AT- binding specifity, but the fluorescence is markedly increased on binding to poly(d(A-T)) (Cavatorta et al. 1985). On the other hand, binding of DAPI to polynucleotides containing guanine and cytosine - although taking place - do not result in an enhancement of fluorescence. Thus, the remarkable increase of DAPI fluorescence quantum yield is AT-selective and is nearly 25-fold compared to the unbound dye. The quantum yield of free DAPI is 3,5% and varies upon binding to DNA between 60% and 92% depending on chromatin structure.

The first ideas on the nature of binding to DNA were rather speculative and still obscure. Some authors postulated intercalative bonds between adjacent base pairs (ski and Szer 1979, Chandra et al. 1977, Chandra and Mildner 1979). Others found evidence for nonintercalative (Williamson and Fennell 1975) and outside bonding to the DNA skeleton (Stepien et al. 1978 and 1979). Sophisticated spectroscopic techniques provide more and more information of the latter type of interaction between DAPI and ds-DNA, establishing the model of selective coupling to the minor groove (Manzini et al. 1983, Kubista et al. 1987, Larsen et al. 1989, Barcellona and Gratton 1990, Tansiou et al. 1992, Mohan and Yathindra 1994, Trotta et al.  2003, Tawar et al. 2003).
From all these studies (performed with native DNA and not fixed as for cytometry) it can be concluded that clusters of four AT base pairs initiate the binding of DAPI, fig.2.


















Fig.2a:
A combined skeletal (Kendrew) and space filling (Corey-Pauling) model of B-DNA
and DAPI bound to the minor groove.
For more info klick here.




























Fig.2b:
Stereo image of the Larsen model of DAPI bound to the B-DNA
decanucleotidyl C-G-C-G-A-A-T-T-C-G-C-G (slightly modified
according to Larsen et al. 1989)

An excellent review on "DAPI: a DNA-specific fluorescent probe" is given by Jan Kapuściński (1995).

In their recent paper Banerjee and Pal (2008) found that both types of bonding may occur:

                                       minor groove binding to AT repeats
                             with pronounced enhancement of fluorescence,
 
                            whereas GC repeats allow intercalative bonding
                              with narrow fluorescence such as the free dye.

In addition, as reported for 33258 Hoechst (Latt and Wholleb 1975) which binds in a similar fashion to DNA (Pjura et al. 1987, Portugal and Waring 1988, Haq et al. 1997, Bostock-Smith and Searle 1999) chromosomal proteins modify the capacity of DNA to associate with DAPI. They may block at least half of the binding sites which are available in native DNA (Darzynkiewicz et al. 1984, Rundquist 1993).
The geometry and electrostatic outfit of the minor groove may vary due to local factors from one base pair to the next (Larsen et al. 1989).
Thus, upon the sensitivity to sequence makeup and unpredictable interaction of proteins, it is rather insecure to calculate the accurate number of binding sites. It might be much smaller than any computation will point out.

The formation of DNA-DAPI complexes and subsequent excitation of fluorescence yield a couple of pro- cesses with their corresponding rate constants. The rate constant is the formation change per time (commonly seconds). The reziprocal rate constant is the transition time.

1. Formation of the DNA-dye complex and its reversed dissociation with the corresponding reaction
    rates.
2. After irradiation the complex is transferred into an excited state and besides fluorescence various
    processes of quenching may occur, each linked to its own rate constant.
3. Photolysis and other disintegrating reactions may happen.

Finally up to eleven reaction constants may play a role in processing the energy of each absorbed
photon (Watson 1991). If all rate constants are stable in itself, than the outcome of fluorescence is
a constant portion of the sum of all processes. In living cells and those which are well preserved by
cross-linking fixatives (glutaraldehyde, formalin and a couple of others which are well established in
histology and cytology, Romeis 1968) cell-to-cell differences in these rate constants are to be ex-
pected, depending on the organisational level and the functional activity of the chromatin (Ringerts
1969
). Despite constancy in DNA content the DAPI fluorescence may vary and the coefficient of variation in DAPI cytometry of cells fixed with aldehydes is 10% and worse (Vinogradov 1995). The crude fixation with ethanol (dehydration) may cut those slight deviations. As a consequence, fixation with methanolic or ethanolic solutions is commonly in use for DNA measurements in flow.
                                           ..................................................
 

  • 2.  The Jablonski diagram

The interaction between light and matter is commonly illustrated with a diagram, mostly exerted itself, which has been invented by Jablonski (1935), fig.3a, an ukrainian physicist and the father of fluorescence spectroscopy. It is given in fig.3b.






Fig.3a:
Alexander Jablonski (1898-1980)
was born in Voskresenovka, Ukraine.
Among other accomplishments,
he defined the term
"anisotropy”,
describing the polarized
fluorescence
from solutions.



















Fig.3B:
A type of Jablonski diagram upon the energy scheme of fluorophores

Basically a three-step process results in fluorescence:
1. a photon is absorbed within 10^-15 sec and raises an electron into a certain vibrational level of a
    higher electronic state which leads to an excited state of the  fluorophore,
2. during internal conversion within 10^-12 sec the electron drops down to the lowest level of the
    first excited electronic state and the corresponding energy is mainly dissipated into heat.
3. A return of the electron within some 10^-9 sec to one of the vibrational levels of the ground state
    leads to emission of a photon.

Due to energy dissipation during internal relaxation of the fluorophore after being excited the energy of the emitted photon is lower than that of the absorbed one. Consequently, the wavelength of absorbed photons is shorter than of the emitted ones. The fluorescence color is shifted into the red compared with the spectrum of the absorbed light. This fact was first (1852) observed by Stokes (fig.4), a Lucasian Professor at Cambridge and is called the “Stokes’ shift”.










Fig.4:

Sir George Gabriel Stokes (1818-1903)











                                           ..................................................
 

  • 3.  Long time fluorescence fading elucidates chromatin fractions differing in the degree of
         condensation

As for any other fluorescent dye the fluorescence of DAPI bound to cellular DNA exhibits pronounced photo-bleaching under prolonged irradiation as shown in fig. 5. The rate of photo-bleaching is charac- teristic for each substrate-dye complex. A mixture of fluorescent complexes differing in their macro- molecular organisation may fade at various rates which are superimposed.



















Fig.5:
Photo-decomposition of DAPI bound to nuclear DNA of diploid liver cells
of the mouse under continuous excitation over minutes.

The decay curve of DNA-DAPI is a superposition of three exponentials with different time constants indicating that three types of divergently structured binding sites are present. Studies on the photo-
decomposition of Feulgen stained nuclei (Benwitz 1965, Haag et al. 1971) have demonstrated, that isopyknotic (euchromatin), heteropyknotic (heterochromatin) chromatin and a third intermediate chromatin fraction (found by Frenster et al. 1963, in biochemical studies) are differently sensitive to photo-bleaching. Haag et al. (1971) applied photo-decomposition for the determination of eu- versus heterochromatin in normal and tumor cells.

Photo-bleaching does not occur at all in the first microseconds of excitation (Bergquist 1973, Rost 1991) under moderate illumination intsensity (50 mWatt and less) and the time constants describing the subsequent fading are rather high.

Therefore photo-bleaching has no significant influence on flow cytometric analyses!
But microscopic studies have to face these circumstances immediately!
                                           ..................................................
 

  • 4.  Short time fluorescence fading exhibits information on binding mechanisms

After irradiation the lifetime of the excited states of any fluorochrome is about some nanoseconds.
A typical fluorescence decay measurement is shown in fig.6.






























Fig.6:
Plot of the excitation pulse and time-resolved fluorescence of R6G
exhibiting a single exponential decay. The white curve indicates
the best fit with evenly distributed residuals.

Einstein’s (1917) formulations of the transition probabilities for atomic absorption and spontaneous emission of a photon has been modified by Perrin (1926), Lewis and Kasha (1945) and finally by Strickler and Berg (1962) for absorbing and fluorescing systems (composite molecules) in solution for compounds with quite high quantum yield, preferably nearly equal to 1, resulting in the expression


including the following terms in particular:

Tfl     = the fluorescence decay time (life time of the excited state) in seconds;
n       = refractive index of the medium;
ε       = wavelength dependent decadic molar extinction coefficient in 1/(mol * cm);
Ú fl    = wavenumber (= reziprocal wavelength (in centimeter)) of the fluorescence maximum in 1/cm;
Ú abs = wavenumber of the absorption maximum.

Fig.7 gives an impression, how the integral term is to be determined.

With the following values we calculate the fluorescence decay time of DAPI bound to in situ DNA:

n        = 1,33
(n^2 = 1,77)
ε(max) = 2,36 * 10^4 at 350 nm or Úabs
Ú fl     = 2,2321 * 10^4
(Ú fl ^3 = 11,1 * 10^12)
Úabs   = 2,8818 * 10^4

ε/Úabs dÚ = ε(max)/Úabs * (Úleft-Úright)
                  = 0,44 * 10^4

1/Tfl = 2,88*1,77*11,1*0,44*10^7
       = 0,25 * 10^9 / sec

which leads to

Tfl = 4 Nanoseconds. 

Fig.7:
Calculation of the fluorescence life time of DAPI bound to DNA in situ.
Illustration for extracting the spectral integral out of the given
corrected spectra of absorption and emission.

Experimentally, the fluorescence decay of DAPI bound to DNA exhibits a triple exponential decay function (fig.8). The life times are 3,9 nsec, 1,8 nsec and 0,1 nsec representing percentages of the total fluorescence of  85%, 12% and 3% (Cavatorta et al.1985, Barcellona and Gratton 1989 + 1990).
The dominant 3,9 nsec is in excellent agreement with the theoretical value of 4 nsec.





























Fig.8:
Fluorescence lifetime data of DAPI-DNA after convolution analysis.
Plots of the remaining residuals after autocorrelation.
Decay curves have been fitted with the following functions:
(top)       single  exponential decay,
(middle)   double exponential decay,
(bottom)  triple   exponential decay delivering the best fit.
(Cavatorta et al.1985)

Neglecting the 3%-portion, these results indicate a two order binding reaction of DAPI to DNA with a dominant first order part of  85%.
In all these experiments the phosphate to dye ratio (P/D) was rather high (around 150 and more). At lower ratios very close binding of DAPI to previously bound DAPI molecules will shift the fluorescence maximum from the blue (460 nm) into the green (540nm). This new band is caused by weak electro- static interaction between these proximate DAPI molecules. Based on these facts the staining concen- trations of DAPI in cytometry should be 1 to 10 µmol and this is the fact in most protocols.
                                           ..................................................
 

  • 5.  How many photons will excite one DAPI molecule in a low power flow cytometer

The effective cross-section (σ) of light absorption of a DAPI molecule can be calculated from the decadic molar extinction coefficient at maximal absorption (van den Engh and Farmer 1992, Lakowicz 2006, pages 58-59).

The relation is the following:  σ = ln(10) 10^3 ε / N (cm^2)

where                                σ is the cross-section,
                                                  ε
is the decadic molar extinction coefficient,
                                        N is Avogadro's number, the number of molecules in one mol;

ln(1o) corrects for the natural logarithm due to the fact that ε is commonly given in decadic logarithms; whereas 10^3 corrects for solvent concentrations given in moles per liter (10^3 cubic centimeter).
With the following values       ε = 23600 per mol and cm (Kapuściński and Skoczylas 1978)
                                        N = 6,2 • 10^23
we get                               σ = 2,3 • 10^3 • 2,3 • 10^4 / 6,2 • 10^23

and finally                           σ = 8,8 • 10^-17 square centimeter
                                                      =
8,8 • 10^-3   square nanometer
                                                      ~ 1                      square Ångström.

We now will try to evaluate the number of photons ready to excite one molecule of DAPI with its small effective cross-section of nearly 1 square Ångström.
       

The energy of a UV photon at wavelength λ=360 nanometer (corresponding frequency ν=0,83*10^15
/ sec)  is                                E=h*ν
with h=6,625*10^-34 Joulesec as Planck's constant.
So we get                                      E=5,5*10^-19 Joule
as the quantum of energy of said UV photon.
So a laser radiating with 10*^-2 Watts delivers 1/5,5*10^-19=1,8*10^16 photons per second.
In a common flow cytometer this laser is focussed down to a spot for excitation of roughly about 20
µm in diameter with an area       a= 314*10^6 square nanometer.

This equals to                         Z=1,8*10^16 / 314*10^6 = 5,73*10^7

photons per square nanometer and second and leads to

                                        N(σ)=5,73*10^7 * 8,8*10^-3 = 5*10^5

as the reservoir of photons for the excitation of one DAPI molecule per second.

The time of intersection in a common flow system is about a few microseconds. If we assume t(exc)=2 microseconds as the excitation time we are getting only

                                        N(exc)= 5*10^5 * 2*10^-6 = 1

photon to interact with DAPI, which is a remarkable low number.

All the assumptions we have made may be tuneable.

But nevertheless the order of magnitude of the number of photons which might hit and excite each bound DAPI molecule will be low in such a way that low laser power will not be sufficient enough for suitable resolution in flow cytometric sorting of chromsomes and other objects with low DNA content.

This is the basic reason why high power lasers are needed for chromosome sorting.
                                           ..................................................

  • For the following considerations we will put together some values and formulas which we do need later:

Avogadro' number..............................................................:

Boltzmann's constant.........................................................:

Room temperatur...............................................................:

Thermokinetic energy.........................................................:

Molecular weight of Rhodamine 6G.......................................:

Molecular weight of DAPI.....................................................:

mean square diffusion path..................................................:
(Feynman et al. 1970 and Berg 1993)

diffusion coefficient of R6G..................................................:
(Rigler et al. 1993)


Additional considerations:

Liter to cubic-nanometer:


Given a solution of molarity one (1 mol) than one molecule of the solute will populate a volume of




One molecule in a solution of molarity X
will occupy a volume of:
                                           ..................................................
 

  • 6.  Few DAPI molecules are surrounded by a predominance of binding sites


The routine molarity of DAPI solutions is about 1 to 10 µmol. The author himself did use 8 µm; so let us take this value  -  the following calculations should be adapted to differing molarities, respectively.

A DAPI solution of 8 µmol contains the following number of molecules in a liter:


With equation 4 we calculate the average volume which is populated by one DAPI molecule in space:




This results in an average distance between adjacent DAPI molecules:

This leads to a maximal distance of 30 nm between a binding site and surrounding dye molecules.
                                         
..................................................
 


  • How many DAPI molecules are in the volume of a human diploid cell nucleus?


The DNA content of a human diploid cell nucleus is about 6 pg with 6 billion base pairs (Altmann and Katz 1976, Rees and Jones 1972).

The amount of DNA (in pg) in a cell nucleus
relates to its volume (in cubic-µm)
according to the following equation:
(Baetcke et al. 1967 and Nagl 1976)


we convert into

inserting DNA=6 pg
we end up with



which finally leads to a nuclear diameter of

This value is in good agreement with calibrated images published by Stöhr et. al. 1969.


Taking equations 3,4 and 5 the number of DAPI molecules present in a volume of a cell nucleus lodging 6 pg DNA is:

A small assembly of 350 thousand DAPI molecules should take over the function to label the huge amount of 6 billion base pairs quantitatively, which is a ratio of 0,00006 to 1.
                                           ..................................................
 

  • The ratio of binding sites to DAPI molecules

Chromatin, the holder of DAPI binding sites, presents itself in different configurations: helical, nucleo- somal and solenoidal (Voet and Voet 1992).



The double helix has cylindrical
dimensions of 2 nm x 3,4 nm and
10 base pairs  perform one turn.






The fundamental components of
nucleosomal DNA strings are
globular elements with 10 nm
in diameter each of it comprising
200 base pairs in 524 cubic-nm.







6 nucleosomes assemble 1 solenoid:
30 nm broad and 10 nm high.
One solenoid has a volume of
9000 cubic-nm including
1200 base pairs.

Assuming a GC/AT-ratio of unity and a random distribution of base pairs than the probability of quadrupletts with identical pairs is about 0,4%. With 6 billion base pairs in a 6 pg cell nucleus we are dealing with the following relations:

Helical level:
                       0,024 billion = 24 million AT base quadrupletts in a volume of 72,4 cubic-µm (eq.6) lodging
                       about 348 thousand molecules of DAPI end up in the ratio of binding sites / DAPI's



Solenoidal level:
                        1200 base pairs per solenoid yield 5 AT quadrupletts in 9 thousand Cubic-nm.
                        Divided by the mean dye free space around each DAPI molecule of 210 thousand
                        Cubic-nm we obtain


Nucleosomal level:
                         200 base pairs implying 0,8 AT quadrupletts in 523 Cubic-nm deliver the relation


                 In conclusion, depending on the level of chromatin condensation the ratio of
                                              binding sites / DAPI ranges from 70 to 320,
                        with a respective cubic spacing of the binding sites from 13 to 8 nm.

These small values of spacial separation are comparable to the critical distance of fluorescein and other fluorescent organic molecules (Wawilow 1925, Förster 1948, Bojarski 1961) below which the fluorescence intensity and fluorescence polarisation declines to zero. Smaller distances, at higher concentrations, lead to the destruction of excitation energy due to Förster resonance energy trans-
fer along the chain of adjacent fluorophores.
Saturation of all these binding sites would end up in a significant decline of fluoresence intensity of DAPI and cytometric investigations would be a risky or even unprofitable undertaking.

By keeping the dye concentration low enough only a portion of bindings sites will be populated by the dye and the majority of sites will be unstained.
On the other hand, enough binding sites will be blocked by protein and will therefore be unaccessible for the dye.

How does this balancing out of quantitative staining work?

Magic words in this context are diffusion and random walk in biology and they will be the issue of the next session.
                                           ..................................................
 

  • 7.  With a least comment we will close this session:

As already calculated (eq. 5) we have 

molecules of DAPI in one ml and 0,024 billion as the theoretical number (on the helical level) of binding sites in a 6 pg cell nucleus. This leads to 20 million cells in one ml offering the same number of binding sites as molecules of DAPI are present in this volume of one ml.
If one does this and tries to separate the stained cells from the remaining solution by using a 0,22 µm filter unit (for sterilizing aqueous solutions) then a cell free solution with unbound DAPI is ready to be applied for a second staining step of cells. This procedure can be repeated several times. After each step the nuclear fluorescence deminishes according to the decrease of DAPI molecules bound in the preceding step. Figure 9 shows the results.

















Fig.9:
Decay of nuclear fluorescence after successive staining with the  
remaining and deminishing dye concentration

During 5 repeats of this dilution process the nuclear fluorescence decreases from step to step by 25%.
Therefore, in the very beginning, the number of binding sites occupied by DAPI was only a quarter of the value calculated.
Let us assume a nucleus of 6 billion base pairs DNA (6pg and a GC/AT ratio of 1), then roughly a quarter of 24 = 6 million binding sites are accessible for DAPI in said nucleus, which is 1000 to 1.

  Is'nt it a marvel to draw conclusions on the whole by measuring only one tenth of a percent.

Basically, DAPI does not bind in a covalent fashion to DNA rather than by electrostatic forces. So the permanent onset of entropy gives rise to perpetual release of DAPI molecules bound to their sites. Other dye molecules in the neighbourhood may replenish them again. A decreasing number of DAPI molecules consequently leads to a deminishing colonization of binding sites and a loss in overall cellular fluorescence.

Therefore we have been dealing with two questions in understanding the cytometrics of nuclear DNA with DAPI:

1. How large is the average life time (or dissociation rate) of DAPI bound to nuclear DNA?

2. How long does it take that a DAPI molecule – maximally 30 nm apart  –  will reach these deserted sites by random thermal migration to make sure that during the measuring time the number of sites populated by DAPI is rather constant?

The answer will be discussed in the section on diffusion.