Spatially resolved derivatization of solid-phase-synthesis beads with fluorescent dendrimers: Creation of localized microdomains

Article abstract of DOI:10.1002/1522-2675(200210)85:10<3532::AID-HLCA3532>3.0.CO;2-F

Second and third generation Newkome-type trifurcated dendrimers, containing either a coumarin or dansyl fluorescent probe at the dendrimer core, have been synthesized and attached to ArgoGel solid-phase-synthesis beads. Subsequent reaction with rhodamine dye shows that the dye can penetrate throughout the beads to acylate the remaining sites. Thus, it is possible to achieve a spatially resolved microdomain for library formation at the core of the dendrimer, primarily on the bead's periphery, and a second microdomain suitable for derivatization by other reagents such as encoding tags and fluorescent sensors.

Full text of DOI:10.1002/1522-2675(200210)85:10<3532::AID-HLCA3532>3.0.CO;2-F

3532
Helvetica Chimica Acta Vol. 85 (2002)
Spatially Resolved Derivatization of Solid-Phase-Synthesis Beads with
Fluorescent Dendrimers: Creation of Localized Microdomains
by Claudia M.Cardona , Stephan H.Jannach , Hao Huang1), Yukiko Itojima1), Roger M.Leblanc ,
and Robert E.Gawley*
Department of Chemistry, and Marine and Freshwater Biomedical Sciences Center, University of Miami,
P. O. Box 249118, Coral Gables, Florida, 33124-0431, USA (phone: ‡13052843279; fax: ‡13056683313;
e-mail: rgawley@miami.edu)
and
Gary A.Baker and Eric B.Brauns
Los Alamos National Laboratory, Michelson Resource, Bioscience Division, P.O. Box 1663, Mail Stop J586,
Los Alamos, NM 87545, USA
Dedicated with congratulations and best wishes to Professor Dieter Seebach on the occasion of his 65th
birthday! Thanks for all you have done for organic chemistry in general, and organic synthesis in particular.
Second and third generation Newkome-type trifurcated dendrimers, containing either a coumarin or dansyl
¾
fluorescent probe at the dendrimer core, have been synthesized and attached to ArgoGel solid-phase-synthesis
beads. Subsequent reaction with rhodamine dye shows that the dye can penetrate throughout the beads to
acylate the remaining sites. Thus, it is possible to achieve a spatially resolved microdomain for library formation
at the core of the dendrimer, primarily on the bead×s periphery, and a second microdomain suitable for
derivatization by other reagents such as encoding tags and fluorescent sensors.
Introduction. Dendrimers are highly branched, monodisperse polymers com-
posed of well-organized structures [1]. Fig. 1,a illustrates a generic, third-generation
trifurcated dendrimer showing the core and peripheral regions. Initially, research on
dendrimers focused on their synthesis and on examination of their physical properties.
More recently, efforts have shifted to finding applications that take advantage of their
novel properties. The dendritic framework isolates sites in the interior of the dendrimer
with characteristic nano-environments [2]. In fact, molecular probes located at the core
of the dendrimer display physical properties that are modulated by their encapsulation
within the periphery (for a few examples, see [3]). On the other hand, the high degree
of branching gives rise to a periphery with many terminal groups, which can be
derivatized for a variety of purposes (for a few examples, see [4]).
In the past several years, dendrimer chemistry has found applications in the field of
solid-phase synthesis. Dendrimers may be synthesized on solid support or attached to a
solid support to increase the loading capacity of a resin (Fig. 1,b) [5]. Similarly, grafted
dendrimers may be employed as heterogeneous catalysts and reagents [5g,h], or chiral
stationary phases for chromatography [5a].
1
)
H. H. on leave from Shanghai Institute of Organic Chemistry, China; Y. I. on leave from Kagoshima
University, Japan.

Helvetica Chimica Acta Vol. 85 (2002)
3533
Fig. 1. a) A generic backbone of a trifurcated dendrimer such as those presented in this paper, showing how the
periphery encapsulates the core. b) Attachment of a bifurcated dendron to active sites on a solid support increases
the number of active sites for solid-phase synthesis or conversion to multiple catalytic sites.
In pioneering work, Seebach et al. have incorporated TADDOL, a versatile chiral
¬
ligand [6], at the core of an aromatic Frechet-type dendrimer suitable for use as a cross-
linking co-polymer in the preparation of polystyrene beads having the dendritic chiral
ligand embedded throughout the solid support (Fig. 2) [7]. After introduction of a
titanium, these heterogeneous catalysts perform extremely well when compared with
their homogeneous counterparts. The chiral ligands BINOL and Salen may be
embedded in a similar fashion, thus establishing the generality of this strategy for
heterogeneous catalysis [8].
Our work is aimed at combinatorial design of dendritic mimics of receptor proteins
and enzymes attached to solid support. To date, we have used −rational× design to
prepare a series of fluorescent sensors for the marine toxin saxitoxin [9], but we are also
¬
Fig. 2. TADDOL Ligands (T), where the aryl groupsare Frechet-type dendrimershaving styryl groupson the
periphery, may be copolymerized to incorporate the TADDOL moiety throughout the polymer resin [7].

3534
Helvetica Chimica Acta Vol. 85 (2002)
interested in encapsulating a sensor within a dendrimer. We reason that the
proteinogenic receptors of xenobiotics such as saxitoxin are enclosed within an (at
least partially) solvent-excluded binding site [10], and that, perhaps, locating a
receptor/sensor in a similarly solvent-excluded dendrimer would enhance binding and/
or selectivity and/or sensitivity. Moreover, similar reasoning could be applied to the
design of enzyme mimics.
To use combinatorial techniques in this effort, we have developed methodology for
the synthesis and attachment of trifurcated Newkome-type dendrimers to solid support.
Herein, we show, by incorporation of a fluorescent dye at the core of both two- and
three-generation dendrimers, by epifluorescence and laser-scanning confocal micro-
scopy, that the dendrimer does not acylate all the active sites in and on the bead, leaving
numerous sites unreacted and, therefore, available for further derivatization (Fig. 3).
Fig. 3. Cartoon representation of attachment of a dye-containing dendrimer to only some of the active sites in a
polymer resin
To obtain a high degree of encapsulation in the smallest number of generations, we
decided to prepare Newkome-type trifurcated dendrons, based on a tris core [11], but
with t-Bu end-groups to maximize spreading in the second- and third-generation. The
encapsulating effect is rapidly observed in the first three generations with trifurcated
dendrimers [2a][3]. The specific dendrimers we targeted to attach to the solid support
are shown in Figs. 4 and 5. Since our ultimate aim is to build combinatorial libraries at
the core of the dendrimer, we wanted the more robust tert-butyl amides at the
periphery instead of tert-butyl esters. For both the second (Fig. 4) and the third-
generation dendrimers (Fig. 5), we wanted to evaluate different dyes for the dif-
ferent types of microscopy used to examine the distribution in the beads. −Coumarin
343× (ˆ2,3,6,7-tetrahydro-11-oxo-1H,5H,11H-[1]benzopyrano[6,7,8-ij]quinolizine-9-carb-
oxylic acid) was selected because it is a visible fluorophore that can be excited with the
488-nm line of the Ar laser of a confocal microscope and is also suitable for use in
epifluorescence microscopy. On the other hand, the dansyl (ˆ 5-(dimethylamino)-
naphthalene-1-sulfonyl) fluorophore is unsuitable for confocal microscopy with a
visible laser, but fluoresces when using epifluorescence microscopy. Also, it is less
susceptible to self-quenching, when several fluorophores are in close proximity because
of its large Stokes× shift (ca. 200 nm) [12].

Helvetica Chimica Acta Vol. 85 (2002)
3535
Fig. 4. A second-generation (G2) Newkome-type dendrimer having a dye at the core
Preparation of Fluorescent Dendrimers. We have already reported a simple
procedure for preparing the first- and second-generation dendrons needed [13].
Scheme 1 shows the synthesis of the second-generation (G2) nona-acid dendrimer
containing a coumarin fluorophore at the core. [(Benzyloxy)carbonyl]amino one-
generation (G1) triacid 1 and amino G1 triester 2 were coupled with 1-ethyl-3-[3-
(dimethylamino)propyl]carbodiimide (EDCI) in 76% yield to afford the [(benzyl-
oxy)carbonyl]amino G2 nona-ester 3. Removal of the (benzyloxy)carbonyl (Cbz)
group by hydrogenolysis affords amino G2 nona-ester 4. Coupling of amine 4 with
−coumarin 343× (5) with O-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexa-
fluorophosphate (HATU) afforded the G2 coumaryl nona-ester 6 in 77% yield, and
removal of the nine t-Bu groups gave the G2 nona-acid 7 in quantitative yield. For
simplicity, 7 will subsequently be illustrated as a set of two concentric circles, denoting
the two generations of the dendrimer, with a cyan −C× for the coumarin at the coumarin
core, and with the nine carboxylic acid groups denoted as shown in the inset at the
bottom of Scheme 1.
In a similar manner, the third-generation (G3) coumarin dendrimer was prepared
as shown in Scheme 2. Triacid 1 and amino G2 nona-ester 4 were coupled to afford the
Cbz G3 27-ester 8 in 56% yield. Hydrogenolysis gave the G3 amine 9 in 71% yield, and
coupling with coumarin 5 afforded the coumaryl G3 27-ester 10 in 88% yield. Removal
of the t-Bu groups was again quantitative and gave the G3 27-acid 11, which will be
written subsequently in cartoon form as three concentric circles and a cyan −C× in the
middle for coumarin, as shown in the inset.

3536
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 5. A third-generation (G3) Newkome-type dendrimer having a dye at the core. Dyes as in Fig. 4.
The dendrimers containing a dansyl dye at the core were synthesized as shown in
Scheme 3. In two steps, G1 amino triester 2 was acylated with dansyl chloride in 81%
yield, and the t-Bu groups removed in 99% yield to afford G1 dansyl triacid 12. This
could be coupled to G1 amino-triester 2 to give the G2 nona-ester 13 in 82% yield.
Removal of the t-Bu groups with HCOOH afforded the G2 nona-acid 14 in 93% yield.
Nona-acid 14 could be coupled with G1 amino triester 2 to give the G3 27-ester 15 in
35% yield. Alternatively, G1 triacid 12 can be coupled with G2 amine 4 to afford 15 in
67% yield. Removal of the 27 t-Bu groups gives the G3 27-acid 16 in 89% yield.
It is interesting to note the different strategies for the synthesis of the coumarin
dendrimers shown in Schemes1 and 2 with the strategy for the dansyl dendrimers in
Scheme 3. Acylation of G1 triester 2 with coumarin 5 could be accomplished easily with

Helvetica Chimica Acta Vol. 85 (2002)
3537
Scheme 1. Synthesis of G2 Coumarin Nona-acid Dendrimer
EDCI, but the resulting G1 coumaryl triester corresponding to 12 did not survive
HCOOH treatment. In contrast, the coumaryl moiety of G2 nona-ester 6 and G3 27-
ester 10 was quite stable to these conditions. This may be an example of the well-known
dendritic effect [2a], whereby the periphery of the dendrimer encapsulates and protects
functional groups at the dendritic core.
To attach the dendrimers to the resin beads, we planned to use carbodiimide
coupling, followed by amidation with (t-Bu)NH2 to introduce the robust, yet bulky

3538
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 2. Synthesis of G3 Coumarin 27-Acid Dendrimer
(t-Bu)NHCOfunctional group at the periphery. As control experiments to test the
viability of functionalizing the periphery with (t-Bu)NH2, we tested the reaction in
solution phase as shown in Scheme 4. EDCI facilitated the coupling of Cbz G1 triacid 1
with (t-Bu)NH2, which proceeded in 90% yield after column chromatography
(Scheme 4,a). Similarly, as shown in Scheme 4,b, EDCI coupling of coumaryl G2
nona-acid 7 proceeded in 59% yield to give the coumaryl G2 nona-amide dendrimer
19a, while diisopropyl carbodiimide (DIC) coupling of Cbz G2 nona-acid 18 afforded
Cbz G2 nona-amide dendrimer 19b, in 78% yield after chromatography. Although the
NMR spectra of these dendrimers looked clean, the mass spectra showed minor

Helvetica Chimica Acta Vol. 85 (2002)
3539
Scheme 3. Synthesis of G2 and G3 Dansyl Nona- and 27-Acid Dendrimers
contamination with a dendrimer where one, and sometimes two, of the terminal
carboxy groups are substituted with an N-acylurea, which resulted from rearrangement
of the active ester.
¾
Modification of ArgoGel-NH2 Beads with Fluorescent Dyes.
ArgoGel
(Argonaut Technologies) beads are poly(ethylene glycol) (PEG) grafted onto

3540
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 4. Synthesis of Nona-amide Dendrimers
polystyrene and allow high loading due to a bifurcated polystyrene graft-linkage. The
−arms× of the PEG grafts, which have the active amine sites at their termini, impart
enough flexibility that normal (i.e., not solids probe) 13C-NMR spectroscopy may be
¾
used to characterize the resin-bound products. ArgoGel is also compatible with
solvents ranging in polarity from H2Oto toluene. In this work, we employed ArgoGel-
NH2 resin with a 0.43-mmol/g loading and an average 164-mm diameter.
The nona-acid G2 coumarin, 7, and dansyl, 14, and the G3 27-acid coumarin, 11, and
¾
dansyl, 16, dendrimers were coupled to the ArgoGel beads under standard peptide-
coupling conditions, followed by amidation with (t-Bu)NH2 to convert the unreacted
active esters on the dendrimer to N-(tert-butyl) amides. These reactions afforded the
G2 dye/dendrimer-bead conjugates 20 and 21, and the G3 analogs 22 and 23

Helvetica Chimica Acta Vol. 85 (2002)
3541
(Scheme 5). In all cases, the stoichiometry of carboxy (on the dendrimer) to amine (on
the resin) was set at 3 :1. Because of the multifunctional nature of the dendrimers, the
stoichiometry of dye to amine was 1:9 and 1:27 of this ratio, leaving the amine in
−molar× excess over the dendrimer. It is not known how many of the carboxy termini of
the dendrimers reacted, hence the subscript −n× in structures 20 23.
¾
Scheme 5. Attachment of G2 and G3 Coumarin and Dansyl Dendrimers to ArgoGel Beads
That both the coupling and the amidation were successful on the solid support is
evident from a comparison of the 13C-NMR spectrum of the independently prepared
(homogeneous) sample of the coumaryl G2 nona-amide dendrimer 19a with the
analogous dendrimer 20 covalently bound to the bead (Fig. 6). In particular, note the
four labeled signals in the spectrum, which belong to the indicated C-atoms in the
second-generation (periphery) of the dendrimer. The chemical shifts are nearly
identical, and the relative intensities suggest that the number of covalent linkages to the
bead, n, is relatively small. In the spectrum of the dendrimer attached to the bead, the
peak marked with a −b× is due to the resin. The small peaks marked with an −x× at 68 and
25 ppm, not seen in the upper spectrum or in the spectrum of the resin, are due to THF
trapped in the resin or the dendrimer. These two peaks are significantly more
prominent in the NMR spectrum of the G3 coumarin dendrimer 22 attached to the
bead (not shown).
Because of the size of the G2 and G3 trifurcated dendrimers, we anticipated that
diffusion through the dendritic matrix would be slowed relative to smaller molecules
¾
(even if they would fit!?), such that some of the active amine sites of the ArgoGel
resin, especially toward the middle of the bead, may not be accessible. To probe the
distribution of possible unreactive sites, we selected beads modified with G2 and G3
coumarin dendrimers for further derivatization. For this, we selected rhodamine B, a

3542
Helvetica Chimica Acta Vol. 85 (2002)
¾
Fig. 6. a) 13C-NMR Spectrum of 19a. S ˆ CDCl3. b) 13C-NMR Spectrum of ArgoGel beadswith G2 coumaryl
dendrimer covalently attached; 20 (Scheme 5 and text). S ˆ CDCl3; b ˆ minor signal from resin; P ˆ CH2OC-
atoms of polyethylene glycol (PEG) grafts on the resin; x ˆ THF.
dye that fluoresces red instead of the blue-green of the coumarin dye. Moreover, the
absorption maxima of these two dyes are sufficiently far apart (ca. 440 nm for coumarin
and ca. 540 nm for rhodamine) that the two dyes can be excited selectively with
appropriate filters in an epifluorescence microscope.
The phenomenon of fluorescence quenching in epifluorescence microscopy has
been addressed in both whole beads [12] and slices made with a microtome [14], as well
as in confocal microscopy [14]. We did not anticipate that this phenomenon would be
important, when the dye was encapsulated in a dendrimer, but in the absence of the
dendrimer, we wanted to avoid the issue. Therefore, we diluted the rhodamine B
isothiocyanate reagent with 27 equiv. of cyclohexyl isothiocyanate. As shown in
Scheme 6, the G2 coumarin-modified bead, 20, and the G3 coumarin-modified bead,
22, were treated with a mixture of isothiocyanates to afford triply derivatized resins 24
and 25, respectively. The distribution of the two dyes throughout the beads were
determined by epifluorescence microscopy as described in a later section. In the
structures of 24 and 25, −n× has the same meaning as it has in Scheme 5: the number of

Helvetica Chimica Acta Vol. 85 (2002)
3543
Scheme 6. Attachment of Rhodamine Dye to G2 and G3 Coumaryl Dendrimer Beads
dendritic termini covalently linked to the resin. The ratio of cyclohexyl to rhodamine
modified sites is 27:1, but the ratio of dendrimer sites to the cyclohexyl/rhodamine sites
is not known.
As controls for the microscopy experiments described below, we also prepared
beads containing the coumarin and dansyl fluorophores, without the dendrons, in order
to visualize the distribution of the active sites throughout the beads and to test for any
self-quenching between proximal fluorophores in or on the bead. As depicted in
Scheme 7, two sets of beads were modified with only the fluorophores, either coumarin
5 or dansyl chloride, to label every available amino group throughout the bead (i.e., 26
and 27). We also prepared beads with coumarin 5 and cyclohexanecarboxylic acid in
1:9 and 1:27 proportions, 28 and 29, respectively. The role of the cyclohexanecarbox-
ylic acid, a non-fluorescent molecule, is to dilute the fluorophore throughout the resin
to prevent self-quenching by proximal fluorophores on solid support [12][14][15].
Note that all the beads were derivatized under very similar conditions. The only

3544
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 7. Attachment of Pure and Diluted Dyesto ArgoGel Beads
¾
difference was the solvent employed, since coumarin 5 is very soluble in CH2Cl2 and
only slightly in DMF, and the resin swells similarly in both solvents2).
Fluorescence Studies of Dendritic Beads. Three types of microscopy were used in
this work. Laser-scanning confocal microscopy (LSCM), which focuses a laser on a
plane that penetrates intact resin beads, was used to semiquantitatively assess the
distribution of coumarin fluorophores throughout the interior of beads. Because of the
wavelength of the laser (488 nm), this technique was limited to beads containing the
coumarin fluorophore. The colors seen in the LSCM images are false color images that
reflect the intensity of emitted light: red is brightest and blue is the most dim.
Epifluorescence microscopy (EFM) was used to evaluate the distribution of coumarin,
dansyl, and rhodamine fluorophores on beads after they had been physically sliced into
wafers ca. 3 5-mm thick. Environmental-scanning electron microscopy (ESEM) was
used to evaluate the morphology of whole beads and selected wafers. The ESEM
images were taken in a water vapor atmosphere at low vacuum. The surfaces have not
been modified by sputtering of a heavy metal.
¾
Fig. 7 shows LSCM and ESEM images of clusters of ArgoGel beads modified with
the G3 coumarin/dendrimer-bead conjugate (i.e., 22). In the LSCM image on the left,
which represents a plane slicing through the center of the beads, note the uneven
distribution of emitted light throughout the beads, as well as localized spots of more
intense (red) emission on the periphery of some of the beads. In the ESEM image on
the right, note that all of the beads show defects in the surface, including craters and, in
some cases, pieces that have been gouged from the surface but are still attached3). We
¾
2
3
)
)
Argonaut Technologies product specification for ArgoGel .
Such defects are the result of the manufacturing process. Beads taken straight from the jar contain similar
deformities.

Helvetica Chimica Acta Vol. 85 (2002)
3545
Fig. 7. a) LSCM Image (10 Â objective) of a plane bisecting a group of G3 dendrimer beads 22 (false color
shows relative intensity of emitted light: red is brightest). b) ESEM Image of 22.
believe that the bright red spots in the LSCM image correspond to the type of defect
shown in the ESEM image.
LSCM can acquire successive −optical slices× across successive planes in individual
beads, and, therefore, provide a −z-scan× that maps fluorescence intensity in planes
dissecting the bead at different −latitudes×. Fig. 8 shows such a z-scan for bead 22. At this
magnification (40 Â ), each optical slice corresponds to a thickness of ca. 1 2 mm. Note
that the slices nearest the middle of the bead, A and B, are similar to the images in
Fig. 7, and show high concentrations of fluorophore near the periphery and lower
concentrations in the interior. Successive slices show varying distribution patterns of
fluorophore. Near the top of the bead, slices C and D, the fluorophore appears to be
more evenly distributed across each slice.
Fig. 9,a shows the LSCM image of another coumarin/dendrimer bead 22; the
diameter of the image (black line), 166 mm, corresponds closely to the diameter of the
bead, and, therefore, this optical slice is near the center of the bead. In this image, there
is clearly a higher concentration of the dendritic fluorophore near the edge of the bead,
and significantly less emission from the central region, although the distribution in the
middle is far from homogeneous. Defects in the bead are discernable at the top and
right sides of the bead. Fig. 9,b shows a semiquantitative profile of intensity of emitted

3546
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 8. LSCM z-Scan of G3 coumaryl-dendrimer bead 22. The figure in the center shows the direction of view of
the LSCM image, and the black lines indicate the dissecting planes corresponding to the four images shown.
light recorded along the path indicated by the white arrow in Fig. 9,a and clearly shows
higher intensity at the bead×s edge. Fig. 9,c and d show EFM and ESEM images of a
cluster of three wafers, ca. 5-mm thick, cut from three G3 coumarin/dendrimer beads,
22, that were stuck together. In the image in Fig. 9,c, the color is real, and corresponds
to the fluorescence of the coumarin dye at the core of the attached dendrimer. Again,
higher concentrations near the periphery of each wafer are evident. The ESEM image
in Fig. 9,d, of the same three wafers, shows how slicing by a microtome produces a
rather uneven surface when cutting through the polystyrene matrix4), with a tear
evident in the lower left, and holes in two of the wafers. In Fig. 9,e, a wafer cut from a
G3 dansyl/dendrimer bead, 23, shows a layer of thin, −crusty× emission at the edge of the
bead, reminiscent of the solar corona, but otherwise even distribution of fluorophore
across the wafer. Control experiments (vide infra) show that this −corona× is an artifact,
and is not indicative of a microdomain. With the dansyl fluorophore within the
dendrimer, the emission intensity is considerably less, and long exposures (up to several
seconds) were necessary to record an image.
Similar experiments were performed with the G2 coumarin/dendrimer, 20, and G2
dansyl/dendrimer, 21, beads. Fig. 10,a shows a LSCM optical slice near the center of a
bead, and Fig. 10,b shows the profile of emission intensity taken along the direction of
the white arrow. The similarity of these images to the images of the G3 dendrimer beads
(Fig. 9,a and b) is striking. It would appear from these images that, again, the
dendrimer is not distributed uniformly throughout the beads. However, an EFM
photograph of a 5-mm wafer, Fig. 10,c shows considerably less contrast between the
interior and the periphery than the EFM images in Fig. 9,c. EFM Photographs of the
4
)
The wafers sliced by the microtome are far from uniform. See Exper. Part.

Helvetica Chimica Acta Vol. 85 (2002)
3547
Fig. 9. a) LSCM Image of the center of G3 coumarin/dendrimer bead 22. b) Profile showing intensity of emitted
light along path of white arrow in a. c) EFM Image of three microtome wafers, ca. 5-mm thick, of G3 coumarin/
dendrimer bead 22. d) ESEM Image of the same wafers in c. Note the curled corrugated surface and holes in two
of the wafers. e) EFM Image of a microtome wafer of G3 dansyl/dendrimer bead 23.
G2 dansyl/dendrimer beads 21 (again requiring a long exposure time) also show little
discernable contrast (Fig. 10,d)5). Taken together, these images suggest that the G2
5
)
In EFM, the entire specimen is irradiated with UV light, and the resulting fluorescence originates from
throughout the specimen, including above and below the plane of focus. This superimposition obscures
structural details that might otherwise be resolved. In addition, for specimens thicker than the depth of
field (such as these wafers), light from out-of-focus planes may create diffuse halos around objects of
study, degrading contrast.

3548
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 10. a) LSCM Image of the center of G2 coumarin/dendrimer bead 20. b) Intensity profile of emitted light
along path of white arrow in a. c) EFM Image of a wafer, ca. 5-mm thick, of G2 coumarin/dendrimer bead 20.
d) EFM Image of a microtome wafer of G2 dansyl/dendrimer bead 21.
dendrimer penetrates the bead more effectively than the G3. The higher concentration
of fluorophore at the periphery evident in the LSCM image may be due, in part, to a
higher concentration of active sites on and near the bead×s exterior (see also the control
experiments below).
The data presented so far strongly suggest that the dendrimer-fluorophores do not
occupy all the sites in the bead. Indeed, it appears that the G3 dendrimers attach
selectively on or near the periphery of the bead. Since the dendrimer is added in a
substoichiometric amount relative to active amine, this makes sense due to the

Helvetica Chimica Acta Vol. 85 (2002)
3549
presumably slow diffusion rate of the larger dendrimer through the polymer matrix.
The converse of this situation is that there should be numerous unreacted sites, and that
most of them should be in the interior portion of the beads. Thus, beads modified as
outlined in Scheme 6 were sliced and examined for the distribution of coumarin vs.
rhodamine dyes, using filters that selectively excite one fluorophore over the other.
Fig. 11,a shows two images of the same wafer of a G3 coumarin/dendrimer bead that is
also modified with a 1:27 mixture of rhodamine B and a nonfluorescent cyclohexyl
dilutant (25; see Scheme 6)6). In the upper image, fluorescence of the G3 coumarin/
Fig. 11. a) EFM Imagesof a microtome wafer of G3 coumaryl dendrimer/rhodamine bead 25. b) Similar Images
of G2 coumaryl dendrimer/rhodamine bead 24. For both beads, the upper image is taken with a BV filter block:
excitation 400 440 nm, barrier 470 nm. The lower image is taken with a G filter block: excitation 510 560 nm,
barrier 590 nm. See Exper. Part.
6
)
The cyclohexyl dilutant is to prevent proximal quenching of the rhodamine fluorophore in the polymer
matrix [12]. The Stokes× shift of rhodamine is only ca. 20 25 nm, depending on substituents.

3550
Helvetica Chimica Acta Vol. 85 (2002)
dendrimer is clearly visible primarily around the periphery of the wafer, whereas, in the
lower image, the rhodamine is present throughout. The contrasting region in the upper
image is curious, and may be the result of energy transfer from coumarin to rhodamine
chromophores, or −bleeding through× of some rhodamine fluorescence in microdomains
where it is more prevalent. Fig. 11,b shows similar images of the G2 coumarin/
dendrimer bead, modified with rhodamine, 24. In this case, distribution of both
fluorophores is evident throughout the bead.
Thus, with reference to the method of preparation illustrated in Scheme 6, the
rhodamine dye is clearly able to penetrate throughout the resin, past and through both
the G2 and G3 dendrimers. Thus, it is possible to achieve a spatially resolved
microdomain with G3 dendrimers attached to the bead×s periphery, with an interior
microdomain suitable for derivatization by other reagents. Even when diffusion
throughout the bead occurs with the G2 dendrimer, many unreacted sites remain. This
may mean that not all sites are accessible to the G2 dendrimer. In related work, Miller
and co-workers recently demonstrated that beads can be partitioned into as many as
five spatially resolved microdomains by using the kinetics of diffusion of small molecule
reagents into the core of the resin bead [16]. In our work, we made no attempt to
determine the kinetics of diffusion.
Controls. Several papers in the recent literature have addressed the issue of
fluorophores on solid support from a number of perspectives [12][14][17]. For
example, McAlpine et al. used the distribution of fluorophore to evaluate the
distribution of active sites throughout the resin bead [17]7). However, since, to our
¾
knowledge, there are no reports on ArgoGel , we used LSCM and EFM to examine
fluorophore-derivatized beads lacking dendrimers, described above in Scheme 7. Fig. 12
¾
shows a comparison of ArgoGel that has been saturated with 100% coumarin (i.e., 26)
with 10% coumarin/90% cyclohexane (i.e., 28), and with 4% coumarin/96% cyclo-
hexane (i.e., 29). Fig. 12,a shows the all-coumarin bead. The LSCM optical slice at the
top, and the intensity profiles in the middle show clear diminution of emission from the
center of the bead. The EFM image of the wafer at the bottom also shows darkening in
the interior of the wafer. Fig. 12,b shows that, when the fluorophore is diluted with a
derivatizing agent lacking a chromophore, an even distribution of chromophore can be
seen in both the optical slice, its intensity profile, and the microtome slice at the bottom.
Consistent with previous observations [12][14], we believe that the lack of
fluorescence in the center of bead 26 is due to self-quenching by proximal
chromophores, when all sites are occupied. Active sites are closer together in the
center of the bead8). When the fluorophore is diluted such that only 10% of the active
sites have fluorophore, self-quenching is eliminated, and emission is uniform across the
polymer matrix.
Further dilution of fluorophore, as shown in Fig. 12,c, produced a surprise. In
particular, the optical slice now shows a predominance of fluorophore at the periphery
and even distribution of fluorophore throughout the interior. Note that the intensity of
7
)
)
Some of the conclusions of this report have been challenged [14].
On the surface of the bead, active sites are distributed in two dimensions. In the interior, they are
distributed in three dimensions.
8

Helvetica Chimica Acta Vol. 85 (2002)
3551
Fig. 12. Imagesof coumarin-modified beads, asdescribed in Scheme 7. Top : Confocal image through center of
bead. Middle: Profile of intensity of emitted light in confocal image. Bottom: Epifluorescence image of 5-mm
wafer from center of beads. a) All-coumarin bead 26. b) Coumarin/cyclohexyl 1 :9 bead 28. c) Coumarin/
cyclohexyl 1:27 bead 29.
emission, as judged by the noise in the intensity profile, is rather low, and the difference
in intensity is not great. We speculate that there is a sufficiently low concentration of
fluorophore on the bead that stronger emission is only visible where there are high
concentrations of active sites on the resin. In the EFM image at the bottom, a weak
fluorescence emission across the diameter of the wafer is evident; the light −crusty×
edge, reminiscent of the −corona× in Fig. 9,e, is an artifact due to the curling of the edge

3552
Helvetica Chimica Acta Vol. 85 (2002)
producing light refraction5). The EFM image of a wafer cut from the all-dansyl
modified bead 27 (Scheme 7) is indistinguishable from that of 29.
¾
Conclusions. We have shown that ArgoGel beads can be partially derivatized
with G2 or G3 Newkome-type dendrimers, while, at the same time, leaving active sites
within the resin untouched and accessible by other reagents.
Our results show that the confocal and epifluorescence microscopy are useful
techniques for locating functionalized sites on solid-support, given the appropriate
precautions to avoid self quenching of a fluorophore. Derivatization with coumarin
requires dilution to avoid self-quenching, but derivatization with dansyl does not [12].
In our work, the presence of the unmodified amine sites in the dendritically
modified beads produces spatially resolved microdomains that can be differentially
functionalized and visualized by epifluorescence microscopy. Thus, dendrimer-modi-
fied sites can play one role while other sites remain available for other uses such as
encoding [18] or sensing [15].
We are grateful to Dr. John Berry for help with the epifluorescence microscopy, to Dr. Matt Lynn for
running the ESEM experiments, to Mr. Alix Vieux and Dr. Spencer Howell for help with the microtome slicing,
and to Ms. Sara Hinds for assistance with some of the synthesis and chromatography. Financial support from the
US Department of Agriculture (USDA No. 0191270), and core support from the Marine and Freshwater
Biomedical SciencesCenter by NIEHS-NIH (ES 05705) is gratefully acknowledged. C. M. C. thanks NIEHS-NIH
(ES 05931) for an individual fellowship. The contents of this paper are solely the responsibility of the authors
and do not necessarily represent the official views of the NIEHS, NIH. Finally, we are very grateful to Prof. Dr.
Steve Miller and Mr. Gary Copeland (Boston College) for much helpful advice on epifluorescence microscopy.
Experimental Part
General. All commercially available reagents were purchased from Aldrich or Acros, and were used without
further purification unless otherwise noted. THF was distilled from sodium benzophenone ketyl, and CH2Cl2
was distilled from CaH2 or stirred over basic alumina and filtered. Column chromatography (CC) was
performed on silica gel 60 (230 400 mesh), while TLC was carried out using aluminum-backed plates coated
with 0.25-mm silica gel 60 (F-254). UV/VIS Spectra were obtained with a Shimadzu model 2101 double-beam
spectrometer equipped with a cell-holder kept at 258. The spectra are reported as lmax, nm (e, mÀ1cmÀ1). Steady-
state fluorescence spectra were recorded with a Spex Fluoromax system and used uncorrected. For
epifluorescent spectroscopy, see below. NMR Spectra were recorded on 400- or 300-MHz instruments, and
chemical shifts (d) are reported in ppm with either Me4Si or residual solvent as internal reference, as noted. J
values are reported in Hz. Mass spectra (MS) of compounds with a molecular weight below 2000 amu were
recorded by FAB on a VG single-stage quadrupole mass spectrometer. Mass spectra for compounds with a
molecular weight above 2000 amu were obtained by a Bruker MALDI-TOF spectrometer (positive-ion mode;
matrix as indicated), respectively. Atlantic Microlab Inc carried out elemental analyses.
Epifluorescence microscopy was performed with an inverted Nikon TS100-F microscope (10 Â , 20 Â , and
40 Â objectives) with a C-SHG epifluorescence unit and three Nikon filter blocks: a standard UV-1A filter was
used for the dansyl fluorophore (excitation 330 380 nm, barrier 420 nm), standard BV-2A filter for the
coumarin 343 fluorophore (excitation 400 440 nm, barrier 470 nm), and a standard G-2A filter for the
rhodamine B fluorophore (excitation 510 560 nm, barrier 590 nm). Photographs were recorded with a Nikon
CoolPix 990 Digital camera mounted to a trinocular photo tube. Exposure times for the dansyl fluorophore were
1/2 to 2 s in duration. Exposure times for the courmarin 343 fluorophore were 1/1000 to 1/250 s and for the
rhodamine B fluorophore 1/15 to 1/4 s.
Laser-scanning confocal microscopy (LSCM) was performed with a Zeiss Axiovert 100M inverted
microscope coupled to a Zeiss LSM 510 scanning module. Excitation at 488 nm was provided by an air-cooled
argon-ion laser (ca. 10 15 mW). −Survey× scans (e.g., Fig. 7,a) were collected with a 10 Â plan-neofluar
(0.3 NA) objective and subsequent higher-resolution scans (all others) used a 40 Â LD-achroplan (0.6 NA)

Helvetica Chimica Acta Vol. 85 (2002)
3553
microscope objective. In all cases, the coumarin emission was collected by a 488 dichroic mirror and passed
through a 505-nm longpass filter. The typical pinhole setting was varied from 50 100 mm, corresponding to 1.5
2.5-mm thick optical sections. Images were typically acquired in 1 2 s, resulting in a pixel dwell time of 4 8 ms
for the frame size used (512 512). For all samples, the detector gain was adjusted to effectively fill the 8-bit A/D
board.
Scanning electron microscopy (SEM) was performed with a Philips/Electroscan ESEM FEI-XL30 field-
emission-gun environmental SEM. The images of the beads and microtome slices were obtained in a low
vacuum SEM mode with a (1.00 1.40 keV) electron beam. Beads were mounted on a SEM sample holder with
double-stick carbon tape. The microtome slices in Fig. 9,d were mounted on a glass microscope slide.
Syntheses. 2,3,6,7-Tetrahydro-11-oxo-N-{tris[(2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methyl}methyl)-
amino]carbonyl}ethoxy)methyl]methyl}-1H,5H,11H-[1]-benzopyrano[6,7,8-ij]quinolizine-9-carboxamide (6).
Compound 5 (0.10 g, 0.35 mmol) was dissolved in 2 ml of anh. DMF and 5 ml of anh. CH2Cl2. Then, Et3N
(0.12 ml, 0.86 mmol) and HATU [19] (0.14 g, 0.37 mmol) were added, and the mixture was allowed to stir for
5 min under N2. Thereafter, compound 4 [13], dissolved in 5 ml of CH2Cl2, was added, and the mixture was
stirred for 24 h at r.t. Then, the solvent was removed under vacuum, and the residue was purified by CC (SiO2;
AcOEt/hexane 2 :1 to 100 :0). An orange-yellow, viscous oil was obtained (0.58 g, 77%). UV/VIS (MeOH): 437
(3.4  104), 268 (9.4  103). 1H-NMR (400 MHz, CDCl3): 1.44 (s, 27 Me); 2.45 (m, 12 CH2); 2.76 (t, J ˆ 6.0,
CH2); 2.86 (t, J ˆ 6.4, CH2); 3.32 (dd, J ˆ 6.4, 2 CH2); 3.63 (t, J ˆ 6.4, 9 CH2); 3.70 (m, 12 CH2); 3.80 (s, 3 CH2);
6.13 (s, 3 NH); 6.97 (s, NH); 8.52 (s, NH). 13C-NMR (CDCl3, 75 MHz): 20.14; 20.29; 21.22; 27.46; 28.12; 36.18;
37.53; 49.79; 50.21; 59.76; 60.07; 67.04; 67.72; 69.16; 69.39; 80.38; 105.61; 108.25; 110.21; 119.38; 126.95; 147.55;
‡
147.87; 152.67; 162.69; 162.87; 170.86; 170.89. MALDI-MS (anthracene): 2089 ([M ‡ Na] ). Anal. calc. for
C
104H171N5O36 (2067.48): C 60.42, H 8.34; found: C 60.50, H 8.37.
2,3,6,7-Tetrahydro-11-oxo-N-{tris[(2-{[(tris{[2-carboxyethoxy]methyl}methyl)amino]carbonyl}ethoxy)meth-
yl]methyl}-1H,5H,11H-[1]benzopyrano[6,7,8-ij]quinolizine-9-carboxamide (7). Compound
6 (0.28 g,
0.13 mmol) was stirred in 15 ml of 96% HCOOH for 26 h at r.t. The reaction was monitored by TLC, and,
upon completion, the HCOOH was removed under vacuum. Any remaining traces of HCOOH were removed
by dissolving the product in MeOH and evaporating in vacuo. The product was dried under vacuum at r.t. for
18 h. A dark-brown, viscous, and hydroscopic oil was recovered (0.20 g, 99%). 1H-NMR (300 MHz, CD3OD):
1.99 (m, 2 CH2); 2.49 (m, 12 CH2); 2.79 (m, CH2); 2.84 (m, CH2); 3.38 (m, 2 CH2); 3.68 (m, 2 CH2); 3.82
(s, 3 CH2); 7.13 (s, CH); 8.49 (s, CH); 9.27 (s, NH). 13C-NMR (75 MHz, CD3OD): 36.01; 38.44; 61.64; 68.35;
‡
‡
‡
69.06; 70.23; 174.17; 175.85. FAB-MS: 1563 ([M ‡ H] ), 1297 ([M À coumarin] ), 1154 ([M À C16H26NO11] ).
Anal. calc. for C68H99N5O36 ¥ 3 H2O(1616.57): C 50.52, H 6.55; found: C 50.83, H 6.77.
Benzyl N-Tris({2-[({tris[(2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methyl}methyl)amino]carbonyl}-
ethoxy)methyl]methyl}amino)carbonyl]ethoxy}methyl)methylcarbamate (8). Compound
1 [13] (0.058 g,
0.12 mmol) was combined with HOBt (0.016 g, 0.12 mmol) in 3.0 ml of anh. THF, and the mixture was cooled
to 08 in an ice-bath. Then, EDCI (0.079 g, 0.41 mmol) was added followed by Et3N (0.060 ml, 0.41 mmol) and 4
[13] (0.74 g, 0.41 mmol) in 3.0 ml of anh. THF. At this point, 2.0 ml of THF were used to transfer any residual 4
into the reaction flask. The temp. was allowed to slowly rise to r.t., and the mixture was allowed to stir under N2
for 4 d. After removing the solvent under reduced pressure, the residue was dissolved in Et2O, and washed with
0.5m HCl and brine. The org. layers were dried (MgSO4), and the solvent was removed in vacuo. The crude
product was then purified by CC (SiO2; AcOEt/MeOH 100 :0 to 99 :1) to yield a colorless viscous oil (0.41 g,
56%). 1H-NMR (300 MHz, (D6)acetone): 1.46 (s, 81 Me); 2.43 (m, 39 CH2); 3.67 (m, 78 CH2); 5.08 (m, CH2);
6.02 (br. s, NH); 6.63 (br. s, 9 NH); 6.78 (br. s, 3 NH); 7.41 (m, 5 CH). 13C-NMR (100 MHz, (D6)acetone):
28.55; 36.96; 37.91; 60.08; 60.88; 60.96; 68.03; 68.53; 68.76; 69.98; 70.19; 80.63; 128.66; 128.82; 129.37; 171.31;
‡
171.41; 171.52. MALDI-MS (anthracene): 5841 ([M ‡ Na] ). Anal. calc. for C285H497N13O107 (5818.02): C58.84,
H 8.61; found: C 58.76, H 8.79.
Tris({2-[{[tris-({2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methyl}methyl)amino]carbonyl}ethoxy}meth-
yl)methyl]amino}carbonyl]ethoxy}methyl)methylamine (9). Compound 8 (0.23 g, 0.039 mmol) was stirred with
10% Pd/C (0.024 g) in EtOH (10 ml) under H2 at r.t. for 48 h. The mixture was filtered over Celite and through a
0.2-mm Millipore filter. The solvent was removed under reduced pressure at 508, and a colorless oil was
recovered (0.16 g, 71%). 1H-NMR (400 MHz, (D6)acetone): 1.46 (s, 81 Me); 2.45 (m, 39 CH2); 3.35 (s, 3 CH2);
3.67 (m, 75 CH2); 6.66 (br. s, 9 NH); 6.80 (br. s, 3 NH). 13C-NMR (100 MHz, (D6)acetone): 28.55; 36.96; 37.89;
‡
60.69; 68.03; 68.54; 69.89; 69.97; 80.64; 171.43; 171.54. MALDI-MS (anthracene): 5685 ([M ‡ H] ).
2,3,6,7-Tetrahydro-11-oxo-N-{tris([2-({[tris({2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methyl}methyl)amino
]carbonyl}ethoxy}methyl)methyl]amino}carbonyl)ethoxy]methyl)methyl}-1H,5H,11H-[1]benzopyrano[6,7,8-
ij]quinolizine-9-carboxamide (10). Compound 5 (0.0095 g, 0.033 mmol) was dissolved in 5 ml of anh. DMF and

3554
Helvetica Chimica Acta Vol. 85 (2002)
the soln. was cooled to 08 in an ice-bath under N2. Then, Et3N and HATU [19] were added, and the soln. was
allowed to stir for 5 min before the G3 amino 27-tert-butyl ester, 9, dissolved in 2 ml of anh. DMF, was added. An
additional 3 ml of DMF were used to transfer the amine to the reaction vessel. The mixture was allowed to reach
r.t. slowly, and it was stirred for 5 d. Subsequently, the DMF was distilled off under reduced pressure, and the
residue was purified by CC (SiO2; AcOEt/MeOH 100 :0 to 90 :10). A viscous dark orange oil was obtained
(0.15 g, 88%). IR (CH2Cl2): 1727 (CˆO, ester), 1675 (CˆO, amide). 1H-NMR (300 MHz, CDCl3): 1.39
(s, 81 Me); 1.73 (m, 2 CH2); 2.38 (m, 39 CH2); 2.70 (m, CH2); 2.80 (m, CH2); 3.25 (m, 2 CH2); 3.63 (m, 78 CH2);
5.95 (br. s, 3 NH); 6.05 (br. s, 9 NH); 6.94 (s, CH); 8.46 (s, CH); 9.03 (br. s, NH). 13C-NMR (75 MHz, CDCl3):
28.14; 36.17; 37.29; 59.81; 67.06; 67.59; 69.16; 80.40; 170.40; 170.78; 170.89. MALDI-MS (anthracene): 5948
‡
‡
([M ‡ H] ), 2193 ([M ‡ Na À 2 C92H164N4O35] ). Anal. calc. for C293H504N14O108 (5947.44): C 59.13, H 8.54;
found: C 59.08, H 8.50.
2,3,6,7-Tetrahydro-11-oxo-N-{tris([2-({[tris({2-{[(tris{[2-carboxyethoxy]methyl}methyl)amino]carbonyl}-
ethoxy}methyl)methyl]amino}carbonyl)ethoxy]methyl)methyl}-1H,5H,11H-[1]benzopyrano[5,7,8-ij]quinoli-
zine-9-carboxamide (11). Compound 10 (0.11 g, 0.018 mmol) was stirred in 15 ml of 96% HCOOH for 48 h at
r.t. The HCOOH was then removed in vacuo, and any traces of remaining acid were removed by co-distilling
with MeOH under vacuum. The product was allowed to dry at 408 for 8 h in vacuo. A dark-brown, viscous, and
1
extremely hygroscopic oil was recovered (0.081 g, 99%). H-NMR (300 MHz, CD3OD): 1.89 (m, 2 CH2); 2.41
(m, 39 CH2); 2.75 (m, 2 CH2); 3.62 (m, 78 CH2); 7.08 (s, CH); 8.45 (s, CH); 9.17 (s, NH). 13C-NMR (75 MHz,
‡
CD3OD): 36.10; 38.30; 61.70; 68.42; 69.00; 70.26; 174.19; 175.86. MALDI-MS (DHB): 4451 ([M ‡ H ‡ H2O] ),
‡
‡
‡
‡
4469 ([M ‡ H ‡ 2 H2O]) , 4472 ([M ‡ K ‡ H2O]) , 4490 ([M ‡ K ‡ 2 H2O] ), 4508 ([M ‡ K ‡ 2 H2O] ), 4526
‡
‡
‡
([M ‡ K ‡ 3 H2O] ), 3068 ([M À C55H89N4O35] ), 3086 ([M À C55H89N4O35 ‡ H2O] ), 3104 ([M À
‡
‡
C
55H89N4O35 ‡ 2 H2O] ), 3122 ([M À C55H89N4O35 ‡ 3 H2O] ). Anal. calc. for C185H288N14O108 ¥ 8 H2O
(4580.42): C 48.51, H 6.69; found: C 48.57, H 6.67.
5-(Dimethylamino)-N-{tris[(2-carboxyethoxy)methyl]methyl}naphthalene-1-sulfonamide (12). Dansyl
chloride (0.53 g, 2.0 mmol), 2 [13] (0.57 g, 1.1 mmol), and K2CO3 anh. (0.57 g, 4.0 mmol) were stirred in
10 ml of anh. MeCN under N2 at r.t. for 36 h. The mixture was filtered, and the solvent was removed under
reduced pressure. The residue was purified by CC (SiO2; hexane/AcOEt 4 :1) to give slightly green solid (dansyl
G1 triester; 0.73 g, 81%). The dansyl G1 tri-(tert-butyl) ester (0.50 g, 0.67 mmol) was stirred in 8 ml 96%
HCOOH under N2 at r.t. for 24 h. The HCOOH was removed under reduced pressure. Any traces of HCOOH
were removed by dissolving the product in CHCl3 and evaporating in vacuo to give a slightly yellow foam
(0.38 g, 98%). IR (CH2Cl2): 1716 (CˆO). 1H-NMR (300 MHz, (D6)DMSO): 2.13 (t, J ˆ 6.5, 3 CH2); 2.79
(s, 2 Me); 3.23 (t, J ˆ 6.5, 3 CH2); 3.31 (s, 3 CH2); 7.20 8.42 (m, 3 CH2, NH). 13C-NMR (75 MHz, (D6)DMSO):
42.21; 31.36; 59.32; 63.43; 66.09; 112.07; 117.03; 120.53; 124.65; 126.05; 126.42; 136.39; 148.24; 160.22; 169.67.
‡
FAB-MS: 570 (M ).
5-(Dimethylamino)-N-{tris[(2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methyl}methyl)amino]carbonyl}-
ethoxy)methyl]methyl}naphthalene-1-sulfonamide (13). 1-Hydroxybenzotriazole (0.095 g, 0.70 mmol), EDCI
(0.47 g, 2.5 mmol), and Et3N (0.32 ml) were added to a soln. of 12 (0.38 g, 0.66 mmol) in 10 ml of anh. THF.
Then, 5 ml of soln. of 2 [13] (1.20 g, 2.38 mmol) in anh. THF was added, and the mixture was stirred under N2 at
r.t. for 22 h. The solvent was removed under reduced pressure, and the residue was dissolved in Et2Oand washed
with 0.5n HCl and brine. The Et2Olayer was dried (MgSO 4). The solvent was removed under vacuum, and the
residue was purified by CC (SiO2; hexane/AcOEt 1 :2) to give pale blue oil (1.1 g, 82%). IR (CH2Cl2): 1726
(CˆO, ester), 1671 (CˆO, amide). 1H-NMR (300 MHz, (D6)DMSO): 1.36 (s, 81 H, Me); 2.08 (t, J ˆ 6.5,
3 CH2); 2.36 (t, J ˆ 6.2, 9 CH2); 2.80 (s, 2 Me); 3.22 (t, J ˆ 6.5, 3 CH2); 3.32 3.53 (m, 21 CH2); 6.93 (s, 3 NH);
7.19 8.42 (m, 7 NH). 13C-NMR (100 MHz, (D6)DMSO): 28.25; 36.26; 36.63; 45.62; 60.05; 62.78; 67.24; 67.54;
68.83; 69.76; 115.36; 120.48; 123.85; 127.80; 127.90; 129.38; 129.49; 129.82; 139.83; 151.60; 170.60; 170.84.
‡
MALDI-MS: (anthracene) 2033.1 (M ). Anal. calc. for C100H169N5O35S (2032.13): C 59.06, H 8.38; found:
C 59.16, H 8.50.
5-(Dimethylamino)-N-(tris{[2-({tris[(2-carboxyethoxy)methyl]amino}carbonyl)ethoxy]methyl}methyl)-
naphthalene-1-sulfonamide (14). Compound 13 (1.0 g, 0.51 mmol) was stirred in 10 ml of 96% HCOOH under
N2 at r.t. for 20 h. The HCOOH was removed under reduced pressure, and any traces of remaining acid were
removed by co-distilling with CHCl3 under vacuum to give a slightly yellow foam (0.72 g, 93%). IR (CH2Cl2):
1640 (CˆO). 1H-NMR (300 MHz, (D6)DMSO): 2.08 (t, J ˆ 6.4, 3 CH2); 2.40 (t, J ˆ 6.3, 9 CH2); 2.79 (s, 2 Me);
3.22 (t, J ˆ 6.6, 3 CH2); 3.32 3.59 (m, 21 CH2); 6.97 (s, 3 NH); 7.19 8.42 (m, 7 H, NH, CH). 13C-NMR
(75 MHz, (D6)DMSO): 31.72; 33.18; 42.26; 56.78; 59.42; 63.84; 64.17; 65.34; 66.33; 112.01; 117.02; 120.57;
‡
124.52; 124.63; 126.05; 126.40; 136.31; 148.23; 167.36; 169.83. FAB-MS: 1529 ([M ‡ H] ).

Helvetica Chimica Acta Vol. 85 (2002)
3555
5-(Dimethylamino)-N-[tris({2-[({tris[(2-{[(tris{[2-(tert-butoxycarbonyl)ethoxy]methy}methyl)amino]car-
bonyl}ethoxy)methyl]methyl}amino)carbonyl]ethoxy}methyl)methyl]naphthalene-1-sulfonamide (15). 1-Hy-
droxybenzotriazole (0.013 g, 0.098 mmol), EDCI (0.046 g, 0.24 mmol), and Et3N (0.1 ml) were added to a
soln. of 12 (0.039 g, 0.069 mmol) in 10 ml of anh. THF. Then, a 5-ml soln. of 4 [13] (0.50 g, 0.28 mmol) in anh.
THF was added, and the mixture was stirred under N2 at r.t. for 30 h. The solvent was removed under reduced
pressure, and the residue was dissolved in Et2Oand washed with 0.5 n HCl and brine. The Et2Olayer was dried
(MgSO4). The solvent was removed under vacuum, and the residue was purified by CC (SiO2; MeOH/AcOEt
1:50) to give pale blue oil (0.27 g, 67%). IR (CH2Cl2): 1724 (CˆO, ester), 1669 (CˆO, amide). 1H-NMR
(400 MHz, (D6)acetone): 1.33 (s, 27 Me); 2.10 (t, J ˆ 7.3, 3 CH2); 2.32 2.35 (m, 36 CH2); 2.82 (s, 2 Me); 3.27
3.31 (m, 6 CH2); 3.52 3.60 (m, 72 CH2); 6.53 (br. s, 12 NH); 6.75 (br. s, NH); 7.18 8.52 (m, 6 CH). 13C-NMR
(100 MHz, (D6)acetone): 28.84; 37.26; 38.20; 37.73; 46.45; 61.19; 61.30; 63.86; 68.33; 68.83; 66.49; 80.94; 70.15;
70.27; 80.94; 122.06; 124.91; 130.03; 130.49; 131.36; 139.83; 140.01; 141.16; 152.00; 171.85; 171.73. MALDI-MS
‡
(anthracene): 5917.9 (M ). Anal. calc. for C25H34N2O11S (5913.40): C 58.66, H 8.55; found: C 58.38, H 8.79.
5-(Dimethylamino)-N-{tris([2-({[tris({2-[({tris[(2-carboxyethoxy)methyl]methyl}amino)carbonyl]ethoxy}-
methyl)methyl]amino}carbonyl)ethoxy]methyl)methyl}naphthalene-1-sulfonamide (16). Compound 15 (0.17 g,
0.029 mmol) was stirred in 5 ml of 96 % HCOOH under N2 at r.t. for 24 h. The HCOOH was removed under
reduced pressure, and any traces of remaining acid were removed by co-distilling with CHCl3 under vacuum to
give a slightly yellow foam (0.11 g, 89%). IR (CH2Cl2): 1651 (CˆO, ester). 1H-NMR (300 MHz, (D6)acetone):
2.34 (t, J ˆ 2.34, 3 CH2); 2.53 2.58 (m, 36 CH2); 2.90 (s, 2 Me); 3.37 3.52 (m, 6 CH2); 3.63 3.37 (m, 72 CH2);
6.95 (br. s, 12 NH); 7.29 8.57 (m, 7 H, CH, NH); 13C-NMR (75 MHz, (D6)acetone): 31.14; 33.63; 31.14; 41.72;
45.56; 47.75; 46.62; 56.96; 57.05; 59.45; 63.63; 64.24; 65.57; 111.67; 124.79; 126.31; 126.60; 136.01; 148.22; 158.17;
‡
169.15; 169.82. MALDI-MS (a-cyano-4-hydroxycinnamic acid): 4424.8 ([M ‡ Na] ). Anal. calc. for
C
181H286N14O107S ¥ 10 H2O(4579.81): C 47.44, H 6.73; found: C 47.06, H 7.02.
Benzyl N-[Tris({2-[(tert-butylamino)carbonyl]ethoxy}methyl)methyl]carbamate (17). 1-Hydroxybenzo-
triazole (0.14 g, 1.0 mmol), Et3N (0.45 ml, 3.2 mmol), and EDCI (0.63 g, 3.3 mmol) was added to 1 [13] (0.16 g,
0.34 mmol) in 10 ml of anh. DMF. Then, (t-Bu)NH2 (0.15 ml, 1.2 mmol) was added in one portion, and the
mixture was stirred under N2 for 24 h. After removal of the solvent at reduced pressure, the residue was
dissolved in Et2Oand washed with 0.5 m HCl and brine. The Et2Olayer was then dried (MgSO 4), the solvent was
removed in vacuo, and the residue purified by CC (SiO2; AcOEt) to yield a colorless oil, which gradually turned
to white crystals on standing (0.20 g, 90%). 1H-NMR (400 MHz, CDCl3): 1.40 (s, 9 Me); 2.30 (t, 3 CH2); 3.65
(m, 6 CH2); 5.00 (s, CH2); 5.30 (br. s, NH); 5.90 (br. s, 3 NH); 7.27 7.35 (m, 5 CH). 13C-NMR (75 MHz,
CDCl3): 29.19; 38.12; 51.49; 59.18; 66.67; 67.98; 69.36; 128.48; 128.84; 136.91; 155.49; 170.82. FAB-MS: 1274 (2
‡
‡
M ), 637 (M ). Anal. calc. for C33H56N4O8 (636.8): C 62.24, H 8.86; found: C 61.87, H 8.85.
Benzyl N-{Tris[(2-{[(tris{[2-carboxyethoxy]methyl}methyl)amino]carbonyl}ethoxy)methyl]methyl}carb-
amate (18). Compound 3 [13] (1.3 g, 0.69 mmol) was stirred in 15 ml of 96% HCOOH for 18 h. Then,
HCOOH was removed at reduced pressure, and any traces of remaining acid were removed by co-distilling with
CHCl3 and MeOH under vacuum. The product was then dried under reduced pressure at 508 to produce a
colorless oil in quant. yield (0.99 g). 1H-NMR (400 MHz, (D6)acetone): 2.30 (t, J ˆ 6.0, 3 CH2); 2.41 (t, J ˆ 6.0,
9 CH2); 3.56 (t, J ˆ 6.0, 12 CH2); 3.60 (s, 12 CH2); 4.94 (s, CH2); 5.86 (br. s, NH); 6.68 (br. s, 3 NH); 7.06 7.29
(m, 5 CH). 13C-NMR (75 MHz, (D6)acetone): 35.63; 38.28; 60.42; 61.38; 66.89; 68.14; 68.74; 70.17; 70.43;
‡
129.08; 129.21; 129.68; 163.03; 173.02; 173.79. FAB-MS: 1430 ([M ‡ H] ). Anal. calc. for C60H92N4O35 ¥ H2O
(1447.4): C 49.79, H 6.55; found: C 49.85, H 6.79.
2,3,6,7-Tetrahydro-11-oxo-N-(Tris{[2-({[tris({2-[(tert-butylamino)carbonyl]ethoxy}methyl)methyl]amino}-
carbonyl)ethoxy]methyl}methyl)-1H,5H,11H-[1]benzopyrano[6,7,8-ij]quinolizine-9-carboxamide (19a). Com-
pound 7 (0.081 g, 0.052 mmol) was dissolved in anh. DMF (5 ml) under N2, and cooled to 08 in an ice-water bath.
1-Hydroxybenzotriazole (0.067 g, 0.49 mmol) was added followed by EDCI (0.45 g, 2.4 mmol), and (t-Bu)NH2
(0.48 ml, 4.6 mmol). The mixture was allowed to reach r.t. slowly and stirred for 66 h under N2. Then, DMF was
distilled off under reduced pressure at 358. The residue was dissolved in CH2Cl2 and washed with 0.5m HCl and
brine. The org. layer was dried (MgSO4), filtered, and the solvent was evaporated in vacuo. The residue was
washed three times with benzene and dried under vacuum for 2 d, to yield a hygroscopic, brown solid (0.063 g,
59%). 1H-NMR (300 MHz, CDCl3): 1.34 (s, 27 Me); 1.98 (m, 2 CH2); 2.34 (t, J ˆ 7.7, 9 CH2); 2.45 (t, J ˆ 8.0,
3 CH2); 2.80 (m, 2 CH2); 3.33 (m, CH2); 3.69 (m, 21 CH2); 3.86 (s, 3 CH2); 6.48 (br. s, 9 NH); 6.59 (br. s, 3 NH);
7.07 (s, CH); 8.52 (s, CH); 9.18 (br. s, NH). 13C-NMR (75 MHz, CDCl3): 20.14; 37.57; 51.03; 59.86; 60.24; 67.64;
67.82; 69.09; 69.43; 108.15; 109.18; 119.90; 152.71; 163.26; 170.63; 171.32. MALDI-MS (anthracene): 2057.8
‡
t
‡
‡
(M ), 2157.9 ([M À BuNH ‡ C8H18N3O] ), 2257.8 ([M À 2 tBuNH ‡ 2 C8H18N3O] ).

3556
Helvetica Chimica Acta Vol. 85 (2002)
Benzyl N-(Tris{[2-({[tris({2-[(tert-butylamino)carbonyl]ethoxy}methyl)methyl]amino}carbonyl)ethoxy]-
methyl}methyl)carbamate (19b). 1-Hydroxybenzotriazole (0.13 g, 0.99 mmol) and DIC (0.56 ml, 3.6 mmol)
were added to 18 (0.15 g, 0.11 mmol) in 10 ml of anh. THF under N2, and the mixture was stirred for 12 h.
(t-Bu)NH2 (0.35 ml, 3.3 mmol) was then added, and the mixture was stirred under N2 for an additional 24 h.
After removal of the solvent at reduced pressure, the residue was dissolved in Et2Oand washed with 0.5 m HCl
and brine. The Et2Olayer was then dried (MgSO 4), the solvent was removed in vacuo, and the residue purified
by CC (SiO2; AcOEt/MeOH 4 :1) to yield a colorless oil (0.17 g, 78%). 1H-NMR (300 MHz, CDCl3): 1.27
(s, 27 Me); 2.26 (t, J ˆ 5.8, 9 CH2); 2.38 (t, J ˆ 6.0, 3 CH2); 3.55 3.63 (m, 24 CH2); 4.96 (s, CH2); 5.57 (br.
s, NH); 6.39 (br. s, NH); 6.69 (br. s, 3 NH); 7.22 7.28 (m, 5 CH). 13C-NMR (75 MHz, CDCl3): 29.23; 37.53;
37.73; 51.36; 59.38; 60.27; 66.67; 67.97; 68.13; 69.63; 70.22; 128.34; 128.46; 128.87; 130.12; 156.19; 170.63; 169.77.
‡
‡
FAB-MS: 1947 ([M ‡ Na] ), 1925 ([M ‡ H] ).
Coupling the Fluorescent Dendrimers and Dyes to the Resin. A typical procedure used 0.250 g of ArgoGel-
NH2 (0.43 mmol/g), to be swollen in 2 ml of anh. DMF, in a 10-ml fritted polypropylene PD-10 column
(Pharmacia Biotech) used as a reaction vessel. Then, for every equiv. of NH2, 3 equiv. of acid and 3 equiv. of
HOBT were dissolved in 4 ml of DMF. This soln. was added to the resin, followed by 50 equiv. of DIC. The
mixture was slowly rotated in a rotisserie mixer for 5 h at r.t. Then, 10 equiv. of (t-Bu)NH2 were added, and the
mixing was continued for 18 h. At this point, the resin was washed with THF (3Â), MeOH (3Â), H2O(3 Â),
MeOH (3Â), and THF (3Â). Each washing took 20 min. Then, the beads were dried under vacuum for 18 h
over P2O5.
The control beads, 26 29, were prepared in the same fashion, but CH2Cl2 was used instead of DMF because
of solubility problems with 5. These reactions were performed in standard glassware with flea stirrers. The 10%
coumarin resin, 28, was prepared from 5 and cyclohexanecarboxylic acid (3 equiv. of total acid) in a 1:9 ratio,
resp. The 4% diluted coumarin resin, 29, was prepared as described above, but in a 1:27 ratio of coumarin to
cyclohexanecarboxylic acid, respectively.
The G2, 20, and G3, 22, coumarin beads were further modified with rhodamine B to afford 24 and 25. Beads
20 and 22 were swollen in 2 ml of DMF. Then, on the assumption that these beads still had a 0.43 mmol/g loading,
3 equiv. of a DMF soln. of rhodamine B isothiocyanate and cyclohexyl isothiocyanate, in a 1:27 ratio, were
added. The mixture was rotated in a rotisserie mixer for 18 h. Then, the washing and drying process was as
described above.
Procedure for the Microtome Slicing. Beads were embedded in Surgipath Blue Ribbon paraffin and sliced
on a Leica RM 2125 microtome with a Sakura Finetek Accu-Edge high-profile blade. The slices were mounted
on Surgipath Precleaned Snowcoat Xtra 25 mm  75 mm  1.0 mm-glass microscope slides.
Modified beads were distributed evenly on the bottom of a heated (558) 1-cm2 metal mold. The mold was
then filled with melted paraffin at 558 and left standing for 10 min at 558 to allow beads to sink to the bottom of
the mold and the trapped air to escape. If air bubbles were still present, the beads were manipulated with heated
Fig. 13. ESEM Image of a microtome-sliced bead of uneven thickness. Note the large section of spherical surface
in the lower right.

Helvetica Chimica Acta Vol. 85 (2002)
3557
forceps to release the air bubbles. The molds were covered with a microtome mounting plate and placed in a
freezer at À 258 for 15 min. After cooling, the mold was removed, and the plate was immediately mounted on a
Leica microtome equipped with a high-profile blade. Slices of 4 Æ 1-mm thickness were obtained strung into
paraffin ribbons, which were expanded with a 408 water flotation bath. The ribbons were lifted from the bath and
mounted on microscope slides. The slides were placed in an oven at 858 for 15 min to dry the slide and allow the
slices to settle onto the surface of the slide. The paraffin was removed by subsequent treatment with toluene or
xylene (3 Â 10 min). Microtome slicing of the beads was very destructive, and the majority of slices were
partially cut, fragmented, or otherwise severely damaged during the process (Fig. 13). The slices were visually
inspected for uniform thickness and diameter by the addition of a drop of toluene to the slide, so the slices could
be carefully manipulated and then removed and remounted onto another slide for photography or electron
microscopy.
REFERENCES
[1] V. V. Narayanan, G. R. Newkome, Top. Curr. Res. 1998, 197, 19; G. R. Newkome, C. N. Moorefield, F.
Vˆgtle, −Dendritic Macromolecules: Concepts, Synthesis, and Perspectives× VCH, Weinheim, Germany,
1996; G. R. Newkome, C. N. Moorefield, in −Comprehensive Supramolecular Chemistry×, Ed. D. N.
Reinhoudt, Elsevier Sciences, New York, 1996, Vol. 10; D. A. Tomalia, H. D. Durst, Top. Curr. Chem. 1993,
165, 193; D. A. Tomalia, A. D. Naylor, A. Goddard, Angew. Chem., Int. Ed. Engl. 1990, 29,
138.
¬
[2] a) S. Hecht, J. M. J. Frechet, Angew. Chem., Int. Ed. 2001, 40, 74; b) C. B. Gorman, J. C. Smith, Acc. Chem.
Res. 2001, 34, 60; c) C. M. Cardona, S. Mendoza, A. E. Kaifer, Chem. Soc. Rev. 2000, 29, 37; d) D. K. Smith,
F. Diederich, Top. Curr. Chem. 2000, 210, 183; e) A. W. Bosman, H. M. Janssen, E. W. Meijer, Chem. Rev.
1999, 99, 1665; f) M. Fisher, F. Vˆgtle, Angew. Chem., Int. Ed. 1999, 38, 884; g) D. K. Smith, F. Diederich,
Chem. Eur. J. 1998, 4, 1353; h) F. Zeng, S. C. Zimmerman, Chem. Rev. 1997, 97, 1681.
[3] C. M. Cardona, J. Alvarez, A. E. Kaifer, T. D. McCarley, S. Pandey, G. A. Baker, N. J. Bonzagni, F. V.
Bright, J. Am. Chem. Soc. 2000, 122, 6139; C. M. Cardona, T. D. McCarley, A. E. Kaifer, J. Org. Chem.
2000, 65, 1857; F. Vˆgtle, M. Plevoets, M. Nieger, G. C. Azzellini, A. Credi, L. De Cola, V. De Marchis, M.
Venturi, V. Balzani, J. Am. Chem. Soc. 1999, 121, 6290; J. Issberner, F. Vˆgtle, L. De Cola, V. Balzani,
Chem. Eur. J. 1997, 3, 706.
[4] S. E. Stiriba, H. Frey, R. Haag, Angew. Chem., Int. Ed. 2002, 41, 1329; N. Rockendorf, T. K. Lindhorst, Top.
¬
¬
Curr. Chem. 2001, 217, 201; C. M. Casado, I. Cuadrado, M. Moran, B. Alonso, B. GarcÌa, B. Gonzalez, J.
Losada, Coord. Chem. Rev. 1999, 53, 185.
¬
[5] a) F. H. Ling, V. Lu, F. Svec, J. M. J. Frechet, J. Org. Chem. 2002, 67, 1993; b) C. Chi, J. Wu, X. Wang, Z.
Zhao, J. Li, F. Wang, Macromolecules 2001, 34, 3812; c) R. Haag, Chem. Eur. J. 2001, 7, 327; d) S.
Lebreton, S. Monaghan, M. Bradley, Aldrichimica Acta 2001, 34(3), 75; e) A. Basso, B. Evans, N. Pegg, M.
Bradley, Chem. Commun. 2001, 697; f) A. Basso, B. Evans, N. Pegg, M. Bradley, Tetrahedron Lett. 2000, 41,
3763; g) S. C. Bourque, H. Alper, L. E. Manzer, P. Arya, J. Am. Chem. Soc. 2000, 122, 956; h) P. Arya, N. V.
Rao, J. Singkhonrat, H. Alper, C. Bourque, L. E. Manzer, J. Org. Chem. 2000, 65, 1881; i) A. Mahajan, S. R.
Chhabra, W. C. Chan, Tetrahedron Lett. 1999, 40, 4909; j) V. Swali, N. J. Wells, J. Langley, M. Bradley, J.
Org. Chem. 1997, 62, 4902.
[6] D. Seebach, A. K. Beck, A. Heckel, Angew. Chem. 2001, 113, 96; Angew. Chem., Int. Ed. 2001, 40, 92.
[7] H. Sellner, P. B. Rheiner, D. Seebach, Helv. Chim. Acta 2002, 85, 352; P. B. Rheiner, D. Seebach, Chem.
Eur. J. 1999, 5, 3221; H. Sellner, D. Seebach, Angew. Chem., Int. Ed. 1999, 38, 1918; P. B. Rheiner, H.
Sellner, D. Seebach, Helv. Chim. Acta 1997, 80, 2027.
[8] H. Sellner, J. K. Karjalainen, D. Seebach, Chem. Eur. J. 2001, 7, 2873; H. Sellner, C. Faber, P. B. Rheiner,
D. Seebach, Chem. Eur. J. 2000, 6, 3692.
[9] R. E. Gawley, Q. Zhang, P. I. Higgs, S. Wang, R. M. Leblanc, Tetrahedron Lett. 1999, 40, 5461, corrigendum
6135; P. Kele, J. Orbulescu, T. L. Calhoun, R. E. Gawley, R. M. Leblanc, Tetrahedron Lett. 2002, in press;
R. E. Gawley, S. Pinet, C. M. Cardona, P. K. Datta, T. Ren, W. C. Guida, J. Nydick, R. M. Leblanc, 2002,
submitted.
[10] G. M. Lipkind, H. A. Fozzard, Biophys. J. 1994, 66, 1.
[11] G. R. Newkome, X. Lin, Macromolecules 1991, 24, 1443; G. R. Newkome, X. Lin, J. K. Young, Synlett 1992,
53; G. R. Newkome, R. K. Behera, C. N. Moorefield, G. R. Baker, J. Org. Chem. 1991, 56, 7162; C. N.

3558
Helvetica Chimica Acta Vol. 85 (2002)
Moorefield, G. R. Newkome, in −Advances in Dendritic Macromolecules×, Ed. G. R. Newkome, JAI Press,
Greenwich, 1994, Vol. 1.
[12] B. Yan, P. C. Martin, J. J. Lee, Com. Chem. 1999, 1, 78.
[13] C. M. Cardona, R. E. Gawley, J. Org. Chem. 2002, 67, 1411.
[14] J. Rademann, M. Barth, R. Brock, H. J. Egehaaf, G. Jung, Chem. Eur. J. 2001, 7, 3884.
[15] G. T. Copeland, S. J. Miller, J. Am. Chem. Soc. 1999, 121, 4306.
[16] R. A. Farrer, G. T. Copeland, M. J. R. Previte, M. M. Okamoto, S. J. Miller, J. T. Fourkas, J. Am. Chem. Soc.
2002, 124, 1994.
[17] S. R. McAlpine, C. W. Lindsley, J. C. Hodges, D. M. Leonard, G. F. Filzen, J. Comb. Chem. 2001, 3, 1; S. R.
McAlpine, S. L. Schreiber, Chem. Eur. J. 1999, 5, 3528.
[18] M. H. J. Ohlmeyer, R. N. Swanson, L. W. Dillard, J. C. Reader, G. Asouline, R. Kobayashi, M. Wigler,
W. C. Still, Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 10922; H. P. Nestler, P. A. Bartlett, W. C. Still, J. Org.
Chem. 1994, 59, 4723.
[19] L. A. Carpino, H. Imazumi, A. El-Faham, F. J. Ferre, C. Zhang, Y. Lee, B. M. Foxman, P. Henklein, C.
Hanay, C. M¸gge, H. Wenschuh, J. Klose, M. Beyermann, M. Bienert, Angew. Chem., Int. Ed. 2002, 41, 442;
L. A. Carpino, J. Am. Chem. Soc. 1993, 115, 4397.
Received May 31, 2002