The polyhydric alcohols.
`Acyclic polyhydric alcohols.
`By
`E. J. Bourne.
`With 1 figure.
`
`I. Introduction.
`The polyhydric alcohols (polyols) may be divided into two classes, depending
`on whether or not they possess a ring structure. In this section consideration
`is given only to the acyclic polyols, which are known also as glykitols or sugar
`alcohols. The cyclic polyols (cyclitols) were reviewed fully by FLETCHER in 1948,
`and are discussed further in the following section of this volume l .
`The present very extensive knowledge of the glykitols has been accumulating
`since the beginning of the nineteenth century and is based on the firm foundations
`laid by such early workers as BERTHELOT, BERTRAND, LOBRY DE BRUYN and
`EMIL FISCHER. The great interest shown in these substances stems from their
`wide natural distribution (in both the free and combined states), from their close
`relationship to the sugars, from the variety of their chemical reactions and
`from a growing realisation of their industrial potentialities.
`With a few exceptions (e.g. the deoxy-derivatives), the glykitols have the
`molecular formula CnHn+2(OH)n and are usually classified in groups of stereo(cid:173)
`isomers on the basis of the value of n; these groups are named, accordingly,
`tetritols, pentitols, hexitols, etc., the simplest glykitol being the triol, glycerol.
`
`II. Structures and properties.
`A. Structures.
`A glykitol may be prepared conveniently in the laboratory by reduction
`of the corresponding aldose, the potential aldehyde group of the open-chain
`form of the sugar being thus transformed into a primary alcohol group; this
`conversion may be effected by sodium amalgam, by sodium borohydride, by
`catalytic hydrogenation, by electrolytic reduction, by a Cannizzaro-type reaction
`in the presence of Raney nickel, etc. (for details see the review by LOHMAR and
`GOEPP 1949). Reduction of a ketose creates a new asymmetric centre and so
`two glykitols are formed; thus D-fructose yields D-mannitol and D-glucitol.
`The structures of the hexitols and lower homologues shown on page 346 are
`»Titten according to the FISCHER-RoSANOFF convention for expressing the three(cid:173)
`dimensional structures in one plane; the historical development of this con(cid:173)
`vention and its implications have been discussed recently by HUDSON (1948).
`The formulae are derived by laying mechanical models on the plane of the paper
`in such a way that the carbon atoms lie in a straight line and the hydrogen
`and hydroxyl attachments stand above the plane; projections are then taken.
`
`1 See this volume, p. 363ff.: DANGSCHAT, G.: Inosite und verwandte Naturstoffe (Cyclite).
`
`B. Åberg et al. (eds.), Aufbau · Speicherung · Mobilisierung und Umbildung der
`Kohlenhydrate / Formation · Storage · Mobilization and Transformation of Carbohydrates
`© Springer-Verlag oHG. Berlin · Göttingen · Heidelberg 1958
`
`Novo Nordisk Ex. 2028, P. 1
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`346
`
`E. J. BOURNE: The polyhydric alcohols. Acyclic polyhydric alcohols.
`
`It will be recalled that sugars are designated D or L according as to whether they
`can be synthesised from D- or L-glyceraldehyde, by a conventional ascent of
`the series, and not on the basis of their specific rotations; thus the orientation
`of the hydrogen and hydroxyl substituents on the secondary carbon atom ad(cid:173)
`jacent to the primary hydroxyl group determines whether an optical enantio(cid:173)
`morph is designated D or L -
`if in the above convention the hydrogen is on the
`
`Trial
`
`Tetritols
`
`CH20H
`I
`HCOH
`I
`CH 20H
`Glycerol
`
`CH 20H
`I
`HCOH
`I
`HCOH
`I
`CH 20H
`Erythritol
`
`CH 20H
`I
`HOCH
`I
`HCOH
`I
`CH 20H
`D-Threitol
`
`CH 20H
`I
`HCOH
`
`H60H
`I
`HCOH
`I
`HCOH
`I
`CH 20H
`Allitol
`(Allodulcitol)
`
`CH 20H
`I
`HCOH
`I
`HOCH
`I
`HOCH
`I
`HCOH
`I
`CH20H
`Galactitol
`(Dulcitol)
`
`CH 20H
`I
`HCOH
`I
`HOCH
`I
`HCOH
`I
`HCOH
`I
`CH20H
`D-Glucitol
`(Sorbitol)
`(L-Gulitol)
`
`CH 20H
`I
`HOCH
`I
`HCOH
`I
`HCOH
`I
`CH20H
`D-Arabitol
`(D-Lyxitol)
`
`Hexitols
`CH 20H
`I
`HOCH
`I
`HCOH
`I
`HOCH
`I
`HCOH
`I
`CH20H
`D-Iditol
`
`Pentitols
`CH 20H
`I
`HCOR
`I
`HCOH
`I
`HCOH
`I
`CH 20H
`Ribitol
`(Adonitol)
`
`CH 20H
`I
`HOCH
`I
`HOCH
`I
`HCOH
`I
`HCOH
`I
`CH 20H
`D -Mannitol
`
`CH 20H
`I
`HCOH
`I
`HOCH
`I
`HCOH
`I
`CH 20H
`Xylitol
`
`CH 20H
`I
`HOCH
`i
`HOCH
`I
`HOCH
`
`!
`HCOH
`I
`CH 20H
`D-Talitol
`(D.Altritol)
`
`Structures at the Glykitols according to the FISCHER-RoSACS"OFF ConvEntion.
`
`left and the hydroxyl group on the right, the compound is assigned to the D-series.
`This system of nomenclature is unambiguous in the case of simple aldose sugars,
`but difficulties are encountered with the glykitols inasmuch as a primary hydroxyl
`group is present at each end of the carbon chain; for example, the compound
`commonly known as sorbitol could be regarded as a member of both the D -and
`L-series (N.B. It is permissible to turn the planar formulae only in the plane
`of the paper). The difficulty is surmounted by using a name indicative of the
`sugar from which the glykitol can be derived by reduction; in this 'my sorbitol
`is regarded as the trivial name for D-glucitol (syn. L-gulitol) and is used only
`for the natural isomer (LOHMAR and GOEPP).
`It will seem from the structural formulae, which include only the D-member
`of each pair of optical enantiomorphs, that there is only one possible triol, three
`tetritols, four pentitols and ten hexitols, whereas the corresponding figures for
`the aldoses are two, four, eight and sixteen, respectively. This decrease in the
`number of possible structures is due to two reasons: (a) some of the glykitoh;
`(e.g. glycerol, erythritol, ribitol, xylitol, allitol and galactitol) possess a plane
`of symmetry and are, therefore, meso-compounds, and (b) the same glykitol may
`be obtained by reduction of two different aldoses (e.g. D-arabitol from D-arabinose
`and D-Iyxose, D-glucitol from D-glucose and L-gulose, and D-talitol from D-talose
`
`Novo Nordisk Ex. 2028, P. 2
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`Physical properties.
`
`347
`
`and D-altrose). Racemates of glykitols (e.g. DIAhreitol) are known, but they
`usually result from chemical reactions and not from natural sources. It should
`be noted that a derivative of a meso-glykitol, in which substituents are arranged
`unsymmetrically, will exist in enantiomorphic forms, so that a compound of
`this type isolated from a natural source might well display optical activity,
`whereas if the compound were synthesised chemically from the meso-glykitol
`it would be obtained as a mixture of the-D- and L-forms.
`As the homologous series of glykitols is ascended to the heptitols and higher
`members the number of possible isomers continues to increase; there are sixteen
`theoretical structures for a heptitol and most of these compounds are now known,
`although only a few have yet been shown to occur naturally. The system of
`nomenclature follows the same pattern as that given above, the suffix "ose"
`in the corresponding aldose being replaced by "itol" (see Rules of Carbohydrate
`Nomenclature). FISCHER originally used the symbols oc and fJ in his nomen(cid:173)
`clature of the higher sugars, the aldose (I) which arises from D-glucose, by ascent
`CHO
`CH20H
`H~OH
`H~OH
`H~OH
`H~OH
`HO~H
`HO~H
`H~OH
`H~OH
`I
`I
`HCOH
`HCOH
`I
`I
`CH20H
`CH20H
`(I)
`(II)
`
`of the series, being called "oc-D-glucoheptose". As the configurations of these
`substances were established, the need for a less arbitrary system became apparent
`and in 1938 HUDSON devised a new nomenclature, based on his observation
`that the properties of each higher aldose closely resemble those of the hexose
`of like configuration for carbon atoms 1 to 5 (for a review see HUDSON, 1945).
`Compound (I), which has the D-glucose configuration for carbon atoms 3,4,5,6,
`and that of D-gulose for carbon atoms 2, 3, 4, 5, thus became D-gluco-D-gulo(cid:173)
`heptose. Hence the alcohol (II) is gluco-gulo-heptitol; it is a meso-compound.
`
`B. Physical properties.
`With the exception of the lowest member, glycerol, the glykitols are white
`crystalline solids with characteristic melting-points, which have been tabulated
`by PrGMAN and GOEPP (1948) (general), LOHMAR and GOEPP (1949) (hexitols)
`and HUDSON (1945) (heptitols and octitols). D-Glucitol is exceptionally difficult
`to crystallise (ROSE and GOEPP 1939). It does, however, form a relatively in(cid:173)
`soluble pyridine complex (STRAIN 1937), which serves to distinguish it from most
`of the other glykitols; 2-deoxy -D-glucitol forms a similar complex (WOLFROM,
`KONIGSBERG, MOODY and GOEPP 1946).
`The glykitols are soluble in water and, to a smaller extent, in the lower al(cid:173)
`cohols and in acetone; they are usually crystallised from these solvents, or from
`mixtures of them. They are virtually insoluble in most other organic solvents.
`Galactitol is more readily crystallised from water than is glucitol and use is made
`of this fact in the separation of the two hexitols from the mixture obtained
`by hydrolysis and hydrogenation of lactose.
`
`Novo Nordisk Ex. 2028, P. 3
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`348
`
`E. J. BOlTRNE: The polyhydric alcohols. Acyclic polyhydric alcohols.
`
`The best general methods for fractionation of mixtures of glykitols are those
`involving chromatographic procedures. For small-scale analysis, filter-paper
`chromatography is particularly useful (HOUGH 1950); the spots can be detected
`with alkaline silver nitrate (HOUGH), alkaline permanganate (PACSU, MORA and
`KENT 1949), bromocresolpurple in a borate buffer (BRADFIELD and FLOOD 1950),
`etc. Reducing sugars present as impurities would, of course, also be detected
`with these sprays, but they can be distinguished readily from the glykitols by
`means of their colour reaction with aniline trichloroacetate (HOUGH). Good
`separations can be obtained between glykitols having different molecular weights;
`isomeric glykitols are separated less readily. The method can be adapted for
`quantitative work (HIRST and Jmms 1949). Larger-scale fractionations can
`be achieved on columns of Florex XXX (LEW, WOLFROM and GOEPP 1946),
`Silene EF-celite (GEORGES, BOWER and WOLFROM 1946) and cellulose (BARKER,
`BOURNE and CARRINGTON).
`In general, glykitols have low specific optical rotations (e.g. D-glucitol, - 2.0°;
`D-mannitol, -0.2 0 ) and so, when such a compound is isolated, it is often difficult
`to decide whether it has a meso-structure, or not. The question can be solved
`by examination of the acetate, or of the glykitol itself in a borate buffer; in
`each case the optical rotatory power of an enantiomorphic glykitol is usually
`enhanced (BOESEKEN 1949). Other polybasic inorganic oxyacids, at suitable PH
`values, may be used in the same way as boric acid; examples are molybdic acid
`(MERRILL, HASKINS, HANN and HUDSON 1947) and the oxides of antimony and
`arsenic (BADREAU 1921, HATT and HILLIS 1947).
`Infra-red spectra are now proving extremely useful in the comparison of
`carbohydrate samples with authentic specimens (KUHN 1950) and also in struc(cid:173)
`tural determinations (BARKER, BOURNE, STACEY and WHIFFEN 1953, 1954,
`WHISTLER and HOUSE 1953). There can be little doubt that they will be used
`effectively in future studies of glykitols since they already facilitate the detection
`of impurities in certain glykitols produced commercially (BARKER, BOURNE,
`STEPHENS and WHIFFEN).
`
`C. Chemical properties.
`The chemical reactions of glykitols may be divided conveniently into two
`classes: those in which each of the hydroxyl groups reacts as a separate entity
`and those in which two or more hydroxyl groups are jointly involved. The former
`class includes esterifications, etherifications and certain oxidation processes.
`Treatment of a glykitol, under appropriate conditions, with excess of an
`acylating agent, such as acetic anhydride or benzoyl chloride, gives the fully
`esterified product, which is usually readily purified by crystallisation. Of such
`esters, the acetates are most commonly used for characterisations; their physical
`constants have been listed by PIGMAN and GOEPP (1948) (general), LOHl\fAR
`and GOEPP (1949) (hexitol acetates and other derivatives of hexitols) and HUDSON
`(1945) (heptitol and octitol acetates). When the appropriate acid anhydride
`or acid chloride is not available, it is very convenient to use trifluoroacetic an(cid:173)
`hydride to promote a direct reaction between the carboxylic acid and the glykitol
`(BOURNE, STACEY, TATLOW and TEDDER 1949). Preferential esterification of
`the primary hydroxyl groups occurs when a glykitol is treated with a deficiency
`(2 molecular proportions) of benzoyl chloride or tosyl chloride and di-esters,
`prepared in this way, are useful intermediates in the synthesis of many glykitol
`derivatives. However, caution should be exercised in the interpretation of any
`reaction in which a partially-esterified glykitol is employed, particularly if a
`
`Novo Nordisk Ex. 2028, P. 4
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`Chemical properties.
`
`349
`
`trace of alkali is present, because of the known tendency of ester groups to
`migrate from one position to another (c/. PACSU 1945). Tosyl esters of carbo(cid:173)
`hydrates have been reviewed recently by TIPSON (1953).
`Ethers of glykitols may be prepared by the general methods applicable in
`the case of simple aliphatic ethers, although it is often difficult to achieve complete
`etherification; methylation is usually accomplished with methyl iodide and
`silver oxide, methyl sulphate and aqueous alkali, etc. (for a review of these
`methods see BOURNE and PEAT 1950). Preferential etherification of the primary
`alcohol groups is not usually possible, except in the case of triphenylmethyl
`ethers (trityl ethers); trityl chloride in pyridine at room temperature reacts
`far more rapidly with primary alcohol groups than with secondaries, and indeed
`the formation of a trityl ether under these conditions has often been regarded
`as indicative of the presence of a primary alcohol group. Such a conclusion is
`not wholly justified, because cases are known of reaction occurring with secondary
`alcohol groups (HELFERICH 1948). A full review of the relative reactivities of
`hydroxyl groups of carbohydrates has been prepared by SUGIHARA (1953).
`Oxidations of glykitols can be effected by a variety of methods, leading to
`products of different types (PIGMAN and GOEPP 1948). With nitric acid, under
`suitable conditions, the corresponding saccharic acids may be obtained as the
`principal products; thus galactosaccharic acid (mucic acid; galactaric acid) can
`be prepared in good yield from galactitol. However, nitric acid is not specific
`for the conversion of the primary alcohol groups to carboxyl groups and, in its
`place, the more specific oxidants, nitrogen dioxide and oxygen in the presence
`of a platinum catalyst, are now being used increasingly (MEHLTRETTER 1953).
`Other oxidation processes (e.g. bromine water; hydrogen peroxide in the presence
`of ferrous ions; electrolysis) result in the formation of aldoses and ketoses.
`The chemical oxidation of secondary hydroxyl groups in preference to pri(cid:173)
`maries is far more difficult to achieve and usually necessitates the use of pro(cid:173)
`tective substituents. An important example is the transformation of D-glucitol
`into I.-sorbose, an intermediate in the chemical synthesis of vitamin C; this
`entails the conversion of the hexitol into 6-0-benzoyl-l: 3-2: 4-di-0-ethylidene(cid:173)
`D-glucitol (III), oxidation with chromic acid to 1-0-benzoyl-3 : 5-4: 6-di-0-ethyli(cid:173)
`dene-keto-L-sorbose (IV) and removal of the substituents to give I.-sorbose (V)
`(SULLIVAN 1945). Fortunately, biochemical oxidation provides a more direct
`and cheaper alternative and is of great importance industrially. Certain organisms
`of the Acetobacter species (e.g. Acetobacter suboxydans, Acetobacter xylinum) oxidise
`specifically secondary hydroxyl groups, which are adjacent to primaries in glyki(cid:173)
`tols, to keto-groups, thus catalysing such conversions as D-glucitol into L-sorbose,
`D-mannitol into D-fructose and allitol into L-psicose. Detailed studies of the
`factors controlling the oxidation (e.g. species specificity, effect of glykitol con(cid:173)
`figuration) have been made by BERTRAND and others, and have been summarised
`by LOHMAR and GOEPP (1949) (see also page 358).
`The final oxidation process to be discussed brings us to those reactions in
`which two or more hydroxyl groups are involved and concerns the use of glycol(cid:173)
`splitting reagents, of which the two most commonly encountered are lead tetra(cid:173)
`acetate and sodium metaperiodate. The former is usually employed in an organic
`solvent (e.g. glacial acetic acid) and the latter in water. One mol. of oxidant is
`consumed by each IX-glycol group, two by three contiguous hydroxyl groups,
`three by four contiguous hydroxyl groups, etc. As will be seen from VI and VII,
`the products from glykitols are aldehydes and formic acid; when a glycol group
`occurs at the end of a carbon chain (R=H), one of the aldehyde fragments is
`formaldehyde. Thus, by measuring the amount of oxidant consumed and the
`
`Novo Nordisk Ex. 2028, P. 5
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`350
`
`E. J. BOURNE: The polyhydric alcohols. Acyclic polyhydric alcohols.
`
`amounts of formic acid (by titration with alkali) and formaldehyde (by isolation
`as its dimedone derivative) produced, considerable information may be obtained
`concerning the structure of the glykitol. It is advisable, as an additional safe(cid:173)
`guard, to characterise the other fragments in the reaction mixture, as the following
`
`/OiH2
`CH3·HC HCO~
`~06H CH·CH3
`H6o/
`I
`HCOH
`I
`CH20·CO·C6H.
`III
`
`/OiH2
`CH3·HC Hio~
`~OCH CH·CHa
`H6o/
`I
`CO
`I
`CH20·CO·C6H.
`IV
`
`-1-iH2
`H10H
`° HOCH
`H~OH
`I
`-' --COH
`I
`CH20H
`V
`
`II
`
`R
`R
`I
`~HOH
`CHO
`~HOH ~ +
`CHO
`I
`I
`R'
`R'
`
`VI
`
`R
`
`R
`
`6HO
`6HOH
`+
`I
`(CHOH)n ~ nHCOOH
`+
`I
`CHOH
`CHO
`I
`~,
`R'
`
`VII
`
`CHOH
`
`CH20H
`
`CH20
`+ CHO
`Ho6H
`I
`Meo~1
`MeO~H
`MeOCH
`I
`I
`HCOMe ~ HCOMe ~ HCOMe °
`I
`I
`I
`I
`HCOH
`HCOH
`HCOH
`I
`I
`6H2-~
`CH20H
`CH20H
`VIII
`
`example illustrates: 3: 4-Di-O-methyl-D-mannitol contains two cx.-glycol groups
`and might be expected to consume two mols. of sodium metaperiodate, whereas
`it actually consumes only one mol., giving formaldehyde and 2 :3-di-O-methyl-D(cid:173)
`arabinose (see VIII) (BOURNE, HUGGARD, STACEY and TATLOW). The reason
`for this discrepancy is that the arabinose derivative is produced in the aldehydo(cid:173)
`form by fission of one of the two identical glycol groups and then rapidly reverts
`to the pyranose form, which is attacked only slowly by the oxidant, if at all.
`Sodium metaperiodate has been used also for the determination of glykitols
`(RAPPAPORT, REIFER and WIENMANN 1937, RAPPAPORT and REIFER 1937,
`CORCORAN and PAGE 1947, CAMERON, Ross and PERCIVAL 1948). It has been
`employed, for example, by CAMERON, Ross and PERCIVAL in quantitative mea(cid:173)
`surements of the D-mannitol content of seaweeds. An essential feature of the
`method is that the oxidation is allowed to proceed for one minute only; under
`these conditions glucose does not interfere seriously. The reason for the more
`rapid oxidation of the hexitol is that free rotation is possible about the carbon(cid:173)
`carbon bonds, thus enabling the hydroxyl groups to assume favoured dispositions
`
`Novo Nordisk Ex. 2028, P. 6
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`Chemical properties.
`
`351
`
`for attack by the oxidant, whereas the glucose structure is less flexible, due to
`the pyranose ring.
`Most of the cyclic acetals (IX) and ketals (X) of the glykitols are crystalline
`and are prepared frequently for characterisation purposes and occasionally as
`the basis of semi-quantitative determinations. They are formed by condensing
`aldehydes or ketones with dihydric alcohols in the presence of acidic catalysts;
`since this process is reversible, the presence of a dehydrating agent is often
`advantageous. The ketals are usually less stable to aqueous acid than are the
`acetals and both are relatively stable to alkali. The presence of fluorine sub(cid:173)
`stituents considerably enhances the stability; for example, trifluoroisopropylidene
`ketals are extraordinarily resistant to acidic hydrolysis (BOURNE, HUGGARD,
`STACEY and TATLOW). A detailed review of the methods of preparation, struc(cid:173)
`tures, properties and physical constants of acetals and ketals of tetritols, pentitols
`and hexitols was prepared recently by BARKER and BOURNE (1952). This group
`of compounds has long presented a fascinating problem in stereochemistry and
`many studies have been made of the structures of individual members of the
`series. It is found that a given aldehyde may react in very different ways with
`isomeric glykitols; for example, formaldehyde gives the 1: 3-2: 4-5: 6-derivative
`of D-glucitol and the 1: 3-2: 5-4: 6-derivative of D-mannitol, but only the 1: 3-4: 6-
`derivative of galactitol. On the other hand, a given glykitol normally reacts
`according to the same pattern with different aldehydes. In 1944, HANN and
`HUDSON deduced, from known structures, an order of preference for acetal
`formation across different possible pairs of hydroxyl groups, subsequent extensions
`to the list being made by NESS, HANN and HUDSON (1948) and by BARKER and
`BOURNE (1952). As a result, it is now possible to predict the structures of nearly
`all of the known cyclic acetals of glykitols on the basis of the following rules:
`If the Greek letters oc, p and yare used to signify the relative positions of the
`two hydroxyl groups engaged in the cyclisation, and C and T to indicate whether
`these groups are cis or trans in the usual Fischer convention (see above) (C and T
`being necessary only when both groups are secondary), then:
`(1) the first preference is for a pC-ring;
`(2) the second for a poring;
`(3) the third for an oc-, ocTo, pT- or yT-ring;
`(4) in methylenation, a pT-ring takes precedence over an ocT- or yT-ring;
`(5) in benzylidenation and ethylidenation, an ocT-ring takes precedence over
`apT- or yT-ring;
`(6) rules (4) and (5) may not apply when one (or both) of the carbon atoms
`carrying the hydroxyl groups concerned is already part of a ring system.
`It was pointed out by BARKER and BOURNE (1952) that rule (5) was based
`on only a few examples and was to be regarded as tentative, that no case had
`been reported of a cyclic acetal of a glykitol containing a yo, ocC- or yC-ring,
`and that the relative stabilities of cyclic acetals to acidic hydrolysis were in
`accordance with the rules. The rules do not apply to cyclic ketals, for which
`five-membered rings are favoured.
`A study of the principles which determine these selection rules has been made
`by BARKER, BOURNE and WHIFFEN (1952), who demonstrated that the marked
`tendency for a carbon chain to adopt the planar zig-zag form affords an expla(cid:173)
`nation of the main features of the rules, since it appears that the most favoured
`rings involve least energy for distortion of the planar chain. Formulae XI and XII
`represent, respectively, D-talitol and allitol, written according to the FISCHER
`convention, whilst XIII and XIV show the same compounds in their planar
`
`Novo Nordisk Ex. 2028, P. 7
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`352
`
`E. J. BOURNE: The polyhydric alcohols. Acyclic polyhydric alcohols.
`
`zig-zag forms, each being viewed perpendicularly to the plane through the carbon
`atoms. It will be seen that, in XIII and XIV, O"T' Ope and 0YT lie on the same
`side of this plane as does the reference oxygen atom, Or' whilst O"e, OpT and
`Oye lie on the opposite side. The distance between Or and Ope in the zig-zag
`form would be 2.51 A., so that to attain the 0-0 distance of 2.34 A., calculated
`for an acetal, would require very little distortion, which could be achieved without
`departure from tetrahedral angles at the carbon atoms; thus pC-rings should
`be formed very readily. Similar analyses were made for other possible rings
`and the results showed a large measure of agreement with the empirical rules.
`
`R
`
`R /"
`6HO
`=b "'0/
`CHO/ "'R"
`~,
`
`X
`
`IX
`
`OH2·OH
`
`1
`
`OH2·OH
`
`Ho6H
`
`Ho6H
`
`Ho6H
`
`H60H
`
`6H2'OH
`XI
`
`yT 2 yO
`
`{3T 3 {30
`
`r 5 r
`
`6
`
`H60H
`
`H60H
`
`H60H
`
`H60H
`
`6H2'OH
`XII
`
`+
`-
`o represents an oxygen atom projecting towards the observer. 0 represents an oxygen
`atom projecting away from the observer. The hydrogen atoms at positions 2, 3, 4, and 5
`are not shown.
`
`Additional factors which determine the selection rules have been discussed
`by MILLS (1954). In the formation of five-membered acetal rings, which are
`practically planar, structures that have two bulky groups in adjacent cis-positions
`of the ring will be avoided if possible; hence exT- and ex-rings are favoured over
`exC-rings. Six-membered acetal rings are likely to exist in a stable chair con(cid:173)
`formation resembling cyclohexane, and will form preferentially if bulky sub(cid:173)
`stituents are equatorial rather than axial. The preferred pC- and p-rings need
`have no large axial substituents, whereas a pT-ring must have at least one.
`MILLS has shown that conformational analysis is particularly useful in the case
`of polyacetals which contain di- and tri-cyclic systems (these cases were not
`fully covered in the empirical rules-see rule 6).
`Another interesting feature of these compounds is that those formed from
`aldehydes (other than formaldehyde) and from unsymmetrical ketones possess
`one or more asymmetric carbon atoms in excess of those in the parent glykitol;
`
`Novo Nordisk Ex. 2028, P. 8
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`Chemical properties.
`
`353
`
`these are the carbon atoms which were originally in the keto-group of the carbonyl
`compound (compare IX and X). Thus, it might be supposed that a mono-acetal
`would exist in two forms, differing in the orientation of the substituents on the
`carbon atom of the ring derived from the aldehyde group. In practice, only one
`form is usually isolated; in the case of an acetal with a six-membered ring, this
`is presumably the form in which the more bulky of the two substituents occupies
`the equatorial position. Diastereomers of this type were isolated, however, by
`NESS, HAHN and HUDSON (1948) following benzylidenation of perseitol, but the
`less stable form rearranged to the more stable one extremely readily, which
`might explain why earlier workers had failed to isolate similar isomers. More
`recently, BOURNE, HUGGARD, STACEY and TATLOW have obtained what are
`almost certainly stable diastereoisomeric forms of trifluoroisopropylidene deri(cid:173)
`vatives of D-mannitol, interconversion of the two forms presumably being pre(cid:173)
`vented by the extraordinary stability of the ring systems (see page 351).
`Mention has already been made (page 348) of the enhancement of the optical
`rotatory powers of glykitols with borate buffer. This phenomenon is the result
`of complex formation, which has been interpreted by BOESEKEN (1949) as follows:
`
`R
`I
`/OH
`CH-O,
`I 'B
`CH-O/ "'-OH
`I
`R
`
`H+ -(cid:173)+---
`
`monodiol boric acid
`
`R
`I
`CH-O
`I
`)B-OH+H20
`CH-O
`I
`R
`
`neutral
`
`~l-
`
`H+ + 2 H 20
`
`i
`H-O"'-B/O-HCI
`CH-O/ ""'O-HC
`I
`I
`R
`RJ
`bisdiol boric acid
`
`C1
`
`
`
`H++
`
`-(cid:173)+---
`
`He noted that the first of these three equilibria is situated very much to
`the right and that the last is dependent upon the position of the two hydroxyl
`groups in the dio!. Since the bisdiol acid, at the dilution usually involved, is
`entirely split into ions, the hydrogen ion concentration, under otherwise equal
`conditions, is a criterion of the more or less favourable position of the hydroxyl
`groups for the attachment of boric acid. In the presence of a glykitol which
`readily forms a complex (e.g. D-mannitol), boric acid behaves as a strong mono(cid:173)
`basic acid and can be titrated accurately. The formation of charged complexes
`enables glykitols, dissolved in borate buffer, to migrate during paper ionophoresis
`and, in this way, good separations can be achieved between unsubstituted and
`substituted glykitols
`(BARKER, BOURNE, FOSTER, PINKARD and STACEY).
`23
`Handbuch d. Pflanzenphysiologie, Bd. YI.
`
`Novo Nordisk Ex. 2028, P. 9
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`354
`
`E. J. BOURNE: The polyhydric alcohols. Acyclic polyhydric alcohols.
`
`BOESEKEN has summarised the stereochemical factors which control the formation
`of borate complexes of glykitols, but the conclusions reached might now be
`worth reconsideration in the light of the above rules governing acetal formation,
`particularly since FOSTER (1953) has shown that pC-rings playa major role in
`complex formation between sugars and borate (the sugars assume the aldehydo(cid:173)
`form when giving complexes of this particular type).
`There is now a renewed interest in the inner ethers or anhydrides of the
`glykitols, which are produced by acid-catalyzed dehydration at elevated tempe(cid:173)
`ratures and by alkaline hydrolysis of certain glykitol derivatives. These deri(cid:173)
`vatives contain groups, the hydrolysis of which leads to the transitory formation
`of carbonium cations (e.g. halogen, nitric or sulphonic acid esters). Excellent
`reviews of the methods of formation and properties of these interesting compounds
`have been prepared by PEAT (1946) and by WIGGINS (1950).
`
`III. Occurrence and enzymatic synthesis and transformation.
`A. Occurrence.
`Many reports are to be found in the literature of the isolation of glykitols
`from natural sources; they have been discovered in both land and sea plants
`and are substrates, or metabolic products, of a variety of micro-organisms. Their
`isolation usually entails extraction with boiling water, acetone or the lower
`alcohols, concentration of the extracts and crystallisation. It is often necessary
`to introduce additional treatments, such as steam distillation, fermentation,
`chromatography, etc., to remove impurities. In their characterisation, useful
`information may be gained by comparing their infra-red spectra (page 348) and
`their behaviour on paper chromatograms (page 348) and paper ionophoretograms
`(page 353) with those of authentic specimens, but these techniques should not
`be regarded as alternatives to the preparation of crystalline derivatives.
`Methods which have been employed for determination of the glykitol contents
`of living organisms include measurements of periodate consumption (page 349),
`of optical activity in the presence of polybasic inorganic acids (page 348), of copper
`complex formation (BERTRAM and RUTGERS 1938) and of the extent of reaction
`with potassium iodomercurate in alkaline solution (FLEURY and MARQUE 1929).
`All of these methods are unsatisfactory inasmuch as they are not specific and
`it seems probable that they will be replaced entirely by procedures based on
`quantitative chromatographic separations (page 348).
`In spite of the wide occurrence of the glykitols, only a minority of them has
`yet been identified in nature; these include glycerol, erythritol, n-arabitol, ribitol,
`galactitol, n-glucitol, L-iditol, n-mannitol, n-perseitol (n-manno-n-gala-heptitol)
`and n-volemitol (n-manno-n-talo-heptitol). It remains to be seen whether modern
`techniques for the fractionation of complex mixtures will facilitate the isolation
`of others or of the enantiomorphs of those already found. A summary of the
`distribution of the glykitols in the plant kingdom is given in table 1; see also
`BARKER (1955).
`There are, however, a few points arising from table 1 which merit further
`comment. As a source of n-arabitol, the mushroom, Fistulina hepatica, is out(cid:173)
`standing since it contains 8-10% of the pentitol (FREREJACQUE 1939). The
`earlier name for ribitol (adonitol) signified its isolation from Adonis vernalis;
`ribitol occurs also in the Chinese drugs "Chei-Hou" and "Saiko " , which are
`extracted from the roots of certain Umbelliferae.
`Galactitol (dulcitol) was originally called melampyrite after Melampyrum
`nemorosum, from which it was first isolated. It is the principal constituent of
`
`Novo Nordisk Ex. 2028, P. 10
`Mylan Institutional v. Novo Nordisk
`IPR2020-00324
`
`

`

`Occurrence.
`
`355
`
`Table 1. Distribution of glykitols m plants.
`
`Glykitoi
`
`Source
`
`Erythritol
`(meso· Erythritol)
`
`Algae: Trentepohlia iolithuB, umbrina, aurea
`Fungi: Aspergillus niger
`Lichens: Roccella montagnei, tinctoria, juci/ormis, phycopsis; Aspicilia